Ethnography as Method

What Anthropology Sees That Other Disciplines Don’t

Every other discipline in this framework arrives with a model. Operations research arrives with queueing theory and optimization. Public finance arrives with institutional economics and incentive design. Workforce economics arrives with labor markets and human capital. Each model is powerful. Each model is also blind — blind to what the model wasn’t built to see.

Anthropology arrives without a model. It arrives with a method. The method is ethnography: sustained, situated observation of how people actually behave in the contexts where they live and work, combined with the disciplined interpretation of what that behavior means to the people doing it. Ethnography does not begin with a hypothesis to test. It begins with a question to inhabit: What is going on here, and what does it mean to the people involved?

This is the People layer’s foundational claim: you cannot understand a healthcare system — or any human system — by studying its models, its metrics, its org charts, and its policy documents alone. You must also study the people inside it, in their own terms, in situ, over time. The gap between what the system says it does and what the people inside it actually do is not noise. It is the most important signal the system produces.

I. The Anthropological Tradition

Malinowski and Participant Observation

Modern ethnographic method begins with Bronislaw Malinowski’s fieldwork in the Trobriand Islands (1915–1918), published as Argonauts of the Western Pacific (1922). Malinowski established the foundational practice of participant observation: the researcher lives among the people being studied, participates in daily life, learns the language, and observes behavior over an extended period — months or years, not hours or days.

The innovation was methodological and epistemological simultaneously. Methodologically, Malinowski demonstrated that sustained presence produces knowledge that short visits and surveys cannot: the patterns of behavior that only become visible over time, the contradictions between what people say and what they do, the informal systems that operate beneath the formal ones. Epistemologically, he argued that understanding a culture requires grasping it from the inside — what he called “the native’s point of view.” This is not empathy in the colloquial sense. It is a disciplined practice of understanding the categories, logics, and meanings through which people organize their experience.

The relevance to healthcare systems is direct. A compliance audit visits for three days, reviews documentation, and departs with findings. An ethnographer who spent six months inside the organization would produce a fundamentally different account — one that includes the workarounds the staff developed to manage the compliance burden, the informal knowledge networks that compensate for the formal systems’ gaps, the gap between the documented procedures and the actual practices, and the meaning that the staff attach to their work, their organization, and the policy changes they are navigating.

Geertz and Thick Description

Clifford Geertz’s The Interpretation of Cultures (1973) — particularly the essay “Thick Description: Toward an Interpretive Theory of Culture” — provided the epistemological framework for what ethnography produces. Geertz distinguished between thin description (recording observable behavior) and thick description (interpreting the meaning of that behavior within its cultural context).

The famous example: a wink. Thin description records a rapid contraction of the right eyelid. Thick description interprets whether it is an involuntary twitch, a conspiratorial signal, a parody of someone else’s conspiratorial signal, or a rehearsal of a parody. The observable behavior is identical. The meaning is entirely different. Determining which meaning is operative requires understanding the context — the relationships, the norms, the history, the culture within which the behavior occurs.

In healthcare: a clinician overrides a clinical decision support alert. Thin description records the override. The EHR logs it. The quality dashboard counts it. The analytics platform reports a 73% override rate. Thick description asks: why did she override it? Was the alert clinically irrelevant (a sensitivity problem in the algorithm)? Was she alert-fatigued after the fortieth interruption this shift? Did she know something about this patient that the algorithm didn’t? Was she following an informal team norm that overrides are acceptable for this alert type? Was she too rushed to read it? Each interpretation implies a different system problem and a different intervention. The override rate tells you nothing. The meaning of the override tells you everything.

Geertz’s argument is that culture is not a variable to be measured. It is a context within which behavior becomes intelligible. Ethnography’s contribution to the capabilitygraph is precisely this: it makes behavior intelligible in ways that metrics, models, and policy analysis cannot.

The Emic-Etic Distinction

Kenneth Pike’s distinction between emic and etic perspectives (coined in Language in Relation to a Unified Theory of the Structure of Human Behavior, 1954) is foundational to anthropological method and directly relevant to how the People layer relates to the other three layers.

Etic analysis describes a system from the outside, using the analyst’s categories. Operations research describes patient flow using queueing theory. Public finance describes funding using institutional economics. Workforce economics describes labor markets using human capital theory. These are etic frameworks — powerful, generalizable, and analytically precise. They are also outsider frameworks. They describe the system in terms that the people inside it do not use and may not recognize.

Emic analysis describes a system from the inside, using the categories of the people who inhabit it. The compliance officer does not think in terms of institutional isomorphism (DiMaggio and Powell). She thinks in terms of “covering our bases,” “staying off the auditor’s radar,” and “making sure we don’t lose the grant.” The nurse does not think in terms of cognitive load theory (Wickens). She thinks in terms of “I can’t keep track of all this,” “something doesn’t feel right about this patient,” and “I’m just trying to get through this shift.” The emic perspective is not less rigorous than the etic. It is differently rigorous — grounded in the lived experience that the etic models abstract away.

The capabilitygraph needs both. The etic frameworks (OR, public finance, workforce economics) explain why the system behaves as it does at the structural level. The emic perspective (the People layer) explains how the people inside the system experience, interpret, and respond to the structural forces the other layers describe. Neither is complete without the other. A structural analysis that ignores lived experience is a map without terrain. An experiential account that ignores structural forces is a story without explanation.

II. Ethnography of Production Systems

Why This Matters Personally

The ethnographic study of production systems — how people organize labor, resources, and knowledge to produce output within specific institutional and ecological constraints — is a recognized subfield with direct lineage to the capabilitygraph’s concerns.

The anthropology of production encompasses work by Eric Wolf (Europe and the People Without History, 1982) on the integration of local production systems into global political economy, Anna Tsing (The Mushroom at the End of the World, 2015) on supply chains and precarious labor, James Ferguson (The Anti-Politics Machine, 1990) on how development projects impose technocratic models on local systems and produce unintended consequences, and Michael Burawoy (Manufacturing Consent, 1979) on how factory workers actively construct and maintain the social relations of production — not as passive recipients of management directives but as agents who negotiate, resist, and adapt within structural constraints.

Timber production in Southeast Asia — the specific context of fieldwork relevant to this framework — sits at the intersection of these traditions. A timber production system involves: natural resource extraction (ecological constraints), labor organization (workforce dynamics), regulatory frameworks (permits, quotas, environmental law), market dynamics (global commodity prices, supply chains), informal economies (unlicensed logging, local processing), and the lived experience of the workers and communities whose lives are organized around the production system. Studying it ethnographically means understanding not just the material flows and economic structures but how the people involved make sense of their situation, navigate the formal and informal rules, and produce output through practices that the formal models of the system do not capture.

This is precisely what the People layer of the capabilitygraph does for healthcare. A rural health organization is a production system. It produces healthcare. It does so within ecological constraints (population density, geography, disease burden), regulatory frameworks (licensure, compliance, reimbursement rules), market dynamics (payer mix, competition for labor), informal economies (workarounds, unreported practices, off-book arrangements), and the lived experience of the workers and patients whose lives are organized around the system. The ethnographic question is the same: How do the people inside this production system actually produce the output, given the constraints they face, and what does the process mean to them?

III. What Ethnography Produces That Other Methods Don’t

The Gap Between Official and Actual

Every organization has two versions of itself: the official version (the org chart, the policy manual, the strategic plan, the compliance documentation) and the actual version (how work really gets done, by whom, through what relationships and informal arrangements). The gap between these two versions is not a flaw to be corrected. It is a structural feature of all complex organizations, and it is often the source of the organization’s resilience.

Charles Perrow’s Normal Accidents (1984) argued that complex, tightly coupled systems produce accidents as a normal consequence of their complexity — that the interactions between components are too numerous and too nonlinear for any control system to anticipate. Karl Weick’s work on sensemaking in organizations (Sensemaking in Organizations, 1995; and the classic study of the Mann Gulch disaster, “The Collapse of Sensemaking in Organizations,” Administrative Science Quarterly, 1993) showed that organizational resilience depends not on formal procedures but on the capacity of individuals and groups to make sense of ambiguous, rapidly changing situations and to improvise responses.

The ethnographic contribution is to make the gap visible — not as deviation from procedure (the audit perspective) but as adaptive practice (the anthropological perspective). When a nurse develops a personal system for tracking patient status that supplements the EHR — sticky notes on a clipboard, a mental model updated during hallway conversations with colleagues — she is not failing to use the official system. She is compensating for the official system’s limitations with local knowledge and situated practice. The sticky notes are not a compliance problem. They are an organizational resource that the formal system cannot see or value.

Etienne Wenger’s concept of communities of practice (Communities of Practice: Learning, Meaning, and Identity, 1998) provides the theoretical framework: knowledge in organizations is not stored in databases and procedures. It is enacted through practice — through the shared repertoire of routines, stories, tools, and relationships that a group of practitioners develops through sustained engagement with their work. Communities of practice are informal, emergent, and often invisible to management. They are also where the actual work of the organization happens.

Tacit Knowledge and Practical Skill

Michael Polanyi’s concept of tacit knowledge (The Tacit Dimension, 1966) — “we know more than we can tell” — is fundamental to understanding why ethnography produces knowledge that interviews, surveys, and documentation review cannot. Much of what workers know about their work is tacit: embedded in practice, embodied in habit, and accessible only through observation, not through questioning.

An experienced triage nurse in a rural emergency department makes rapid, consequential assessments that draw on thousands of prior patient encounters, pattern recognition that she cannot fully articulate, and contextual knowledge (she knows the patient, knows the family, knows the community’s health patterns). Asked to describe her triage process, she will give an account that is accurate but incomplete — the formal decision algorithm plus “clinical judgment.” The tacit component — the judgment — is visible only to someone who watches her work over time and attends to the cues she responds to, the weight she gives to different information, and the contextual factors she incorporates that no algorithm captures.

Jean Lave and Etienne Wenger’s concept of legitimate peripheral participation (Situated Learning, 1991) explains how tacit knowledge is transmitted: not through instruction but through practice, in the context of a community of practitioners who model, mentor, and gradually grant increasing responsibility to newcomers. The apprenticeship model — which healthcare still uses, through clinical rotations, residency, and preceptorship — is an ethnographically visible knowledge transmission system. Its effectiveness depends on the quality of the community of practice, not on the curriculum document.

Power, Resistance, and Accommodation

Anthropology’s analytical toolkit includes what James Scott called “weapons of the weak” (Weapons of the Weak: Everyday Forms of Peasant Resistance, 1985) — the small, everyday acts of resistance through which subordinate groups protect their interests within systems of power: foot-dragging, false compliance, feigned ignorance, gossip, sabotage. These are not dramatic acts of rebellion. They are the mundane practices through which people who cannot openly challenge the system modify its operation from within.

In healthcare organizations navigating policy change, everyday resistance is pervasive and consequential. The physician who charts the minimum required documentation rather than the thorough note she would prefer — that is resistance to a documentation burden she cannot formally challenge. The compliance officer who interprets an ambiguous regulation conservatively — that is accommodation to a power structure (the auditor) she cannot control. The front-desk staff who develop their own scheduling workarounds because the official scheduling system doesn’t account for the complexity of their patients — that is practical resistance to a system designed for a different context.

These practices are invisible to the etic frameworks. Operations research sees the scheduling outcome (wait times, utilization). Public finance sees the compliance outcome (audit findings, disallowed costs). Workforce economics sees the labor outcome (turnover, vacancy rates). None of them sees the practices that produced those outcomes — the human agency, the meaning-making, the resistance and accommodation that shape how structural forces actually play out in specific organizational contexts.

IV. Ethnographic Method Applied to Healthcare

What an Ethnography of Healthcare Transformation Would Look Like

An ethnography of organizations navigating the RHTP transition would not study policy documents, outcome metrics, or financial models (though it would reference all of these as context). It would study people: the administrators who interpret the grant requirements, the clinicians who adjust their practice to meet new reporting demands, the compliance officers who mediate between federal expectations and organizational reality, the patients who experience changes in how care is delivered, and the community members who observe their local healthcare institution changing in ways they may or may not understand.

The method would involve:

Sustained presence. Not a site visit or an interview protocol but months of embedded observation — attending meetings, sitting in clinics, joining hallway conversations, observing the informal networks through which information and influence flow.

Attention to practice. Not what people say they do (the interview account, the policy document) but what they actually do (the observed behavior, the workaround, the adaptation, the resistance). The gap between the two is where the most important organizational knowledge lives.

Emic categories. Not imposing analytical frameworks from outside (institutional isomorphism, resource dependence, structural inertia) but attending to how the people inside the organization understand and describe their situation. What do they call the transition? How do they talk about it among themselves? What metaphors do they use? What do they fear? What do they hope? What do they joke about?

Thick description. Not counting overrides, measuring turnover, or scoring compliance. Interpreting what the overrides mean, why the turnover is happening, and how compliance is actually achieved in practice — with the contextual richness that makes the interpretation credible and the behavior intelligible.

What Ethnographic Knowledge Enables

Ethnographic knowledge of a healthcare organization enables interventions that structural analysis alone cannot design. If the structural analysis says the organization lacks readiness, the ethnography explains why — and the explanation determines the intervention.

If the organization lacks readiness because its informal leadership (the longtime nurse manager who everyone trusts) is skeptical of RHTP and has been quietly discouraging engagement, the intervention is relational: engage the informal leader, understand her concerns, address them honestly, and enlist her as a champion or at least neutralize her opposition. No amount of grant-writing training will overcome informal leadership resistance.

If the organization lacks readiness because its staff have been through three failed change initiatives in five years and have developed a deep, experientially justified cynicism about organizational transformation promises, the intervention is trust-building: demonstrate credibility through small, visible, delivered-on commitments before asking for large-scale change engagement. No amount of strategic planning will overcome earned cynicism.

If the organization lacks readiness because its community’s relationship with government-funded programs has been historically exploitative — a real and common experience in tribal communities, in communities of color, in communities that have seen programs come and go without lasting benefit — the intervention is fundamentally different from the standard grant-readiness toolkit. It requires cultural humility, community engagement, and the willingness to operate on the community’s terms rather than the funder’s.

None of these insights are available from the structural analysis. All of them are available from ethnographic observation. The People layer exists to produce this kind of knowledge — not as an alternative to the structural layers but as their necessary complement.

V. The Epistemological Position

The People layer makes an epistemological claim that distinguishes it from the other three layers: the people inside a system have knowledge about that system that no external analysis can replicate, and accessing that knowledge requires a method — ethnography — that takes their perspective seriously as a form of evidence.

This is not anti-quantitative. Ethnography does not oppose measurement. It contextualizes measurement — providing the interpretive framework within which metrics become meaningful. A 73% alert override rate is a number. The ethnographic account of why 73% of alerts are overridden — in this organization, by these clinicians, under these conditions, with these meanings — is knowledge. The number triggers investigation. The ethnography directs intervention.

This is not anti-structural. Ethnography does not deny that institutions, markets, and incentive structures shape behavior. It insists that the shaping is not deterministic — that people within structures interpret, navigate, resist, accommodate, and transform the structures they inhabit, and that understanding how they do so is essential to understanding the system.

And this is not soft. The traditions that ground this layer — Malinowski’s participant observation, Geertz’s thick description, Scott’s political analysis of everyday resistance, Burawoy’s ethnography of production — are methodologically rigorous, theoretically sophisticated, and empirically demanding. Ethnographic fieldwork is not “talking to people.” It is a disciplined practice of observation, interpretation, and analysis that produces a specific kind of knowledge: knowledge of how human systems actually function, as experienced by the humans inside them.

The capabilitygraph without this layer can describe healthcare systems, model their dynamics, and predict their structural tendencies. With this layer, it can also understand them — as human enterprises, inhabited by people with purposes, meanings, fears, and agency that no model fully captures but that every intervention must account for.

Intellectual Debts

Malinowski, B. (1922). Argonauts of the Western Pacific. The foundational text of participant observation and ethnographic method.
Pike, K. (1954). Language in Relation to a Unified Theory of the Structure of Human Behavior. The emic-etic distinction.
Polanyi, M. (1966). The Tacit Dimension. Tacit knowledge: “we know more than we can tell.”
Geertz, C. (1973). The Interpretation of Cultures. Thick description and culture as context for meaning.
Burawoy, M. (1979). Manufacturing Consent: Changes in the Labor Process Under Monopoly Capitalism. Ethnography of production and worker agency within structural constraint.
Wolf, E. (1982). Europe and the People Without History. Local production systems integrated into global political economy.
Perrow, C. (1984). Normal Accidents: Living with High-Risk Technologies. Complexity and coupling as sources of systemic failure.
Scott, J. (1985). Weapons of the Weak: Everyday Forms of Peasant Resistance. Everyday resistance as political practice within power structures.
Ferguson, J. (1990). The Anti-Politics Machine: “Development,” Depoliticization, and Bureaucratic Power in Lesotho. How technocratic development projects misread and reshape local systems.
Lave, J. & Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Knowledge transmission through practice in communities of practitioners.
Weick, K. (1993). “The Collapse of Sensemaking in Organizations: The Mann Gulch Disaster.” Administrative Science Quarterly 38(4). Organizational resilience as sensemaking capacity.
Weick, K. (1995). Sensemaking in Organizations. How people in organizations construct meaning from ambiguous events.
Wenger, E. (1998). Communities of Practice: Learning, Meaning, and Identity. Knowledge as enacted through shared practice in informal communities.
Tsing, A. (2015). The Mushroom at the End of the World: On the Possibility of Life in Capitalist Ruins. Supply chains, precarity, and the ethnography of global production.