What is ethnography?

Confusingly, ethnography can be understood as both method and theoretical framework, but also as neither, since an ethnographer may incorporate multiple methods within their ethnographic project and may employ various theoretical frameworks (Savage Reference Savage2000). Cerwonka & Malkki’s (Reference Cerwonka and Malkki2007) classic work, often used to teach ethnographic methods to postgraduate students, illuminates how theoretical frameworks are themselves interrogated, deconstructed and rebuilt in the field.

In recognition of this heterogeneity, I discuss ethnography as an approach. I outline what ethnography entails using examples from the literature to illustrate the unique values of this approach. Ethnography is most closely intertwined with anthropology, but is now used across many disciplines, including sociology, political science, education and health services research. My research background is in anthropology, so this article focuses on anthropology and mental health.

Noticing the things that biomedical research can miss

David Mosse and colleagues, currently engaged in a project ethnographically investigating Open Dialogue interventions for psychosis in the UK, have described why the ethnographic approach is not just useful, but necessary for developing effective clinical interventions. Participant observation ‘involves extended encounters with people, institutions or processes […] so as to allow an everyday experience of the situations under investigation, usually for a year or so’ (Mosse Reference Mosse, Baker and Carroll2023). Writing field notes turns ‘moments into documents so they can be remembered and revisited’ (Mosse Reference Mosse, Baker and Carroll2023). The final step is placing one’s account in historical, social, institutional and structural context in an attempt to develop an interpretation of the meaning of the experiences of the people being studied.

Ethnography cannot test or evaluate an intervention. Instead it can follow what happens when particular people attempt an intervention in a particular place, in particular circumstances. As Mosse et al (Reference Mosse, Baker and Carroll2023) point out, ethnography does not ask whether an approach worked, but rather what happened and how, what it changed, what people made of it and how it felt. Therefore, ethnography’s unique selling point is not to ‘prove the efficacy of an intervention through generalizations’, but to ‘provide knowledge that is necessary to generalize interventions in practice’. Inductive holistic study of the populations and contexts of clinical trials can reveal factors that are important both to the future implementation of a successfully trialled intervention and (perhaps more important) to the understanding of an unsuccessful one. Or, as Armstrong pithily puts it in his recent book: ‘Anthropologists notice things that biomedical researchers miss’ (Armstrong Reference Armstrong2024: p. 2).

For example, Whittle’s (Reference Whittle, Kiely and Millard2024) recent ethnography of a National Health Service (NHS) rehabilitation ward for complex psychosis does not just highlight the importance of relational work by multidisciplinary staff. It also explores institutional factors that facilitate or inhibit this work. Armstrong’s (Reference Armstrong2024) ethnography in a community mental health team illuminates how the way in which services are set up affects not just treatment pathways, but also what clinicians ask about, what they consider important, what they explore or do not explore, emphasise or do not emphasise. Moreover, institutional structures do not just shape clinical practice, but also affect the agency and behaviour of patients, who are themselves aware of the bureaucratic constraints within which clinical teams operate. The bureaucratic regime Armstrong describes does not erase clinical compassion and morality, but directs them along particular paths. Patients, families and other clinicians attempt, more or less skilfully, to navigate these paths to achieve specific goals. Successful treatment of a mental disorder is one of these goals. But there are also others, such as the clinician’s capacity to justify their rationale to institutional organs of oversight: the supervisor, the manager, the General Medical Council, the Care Quality Commission.

Cultural psychiatry

Ethnographic research’s contribution to psychiatry is not new. Classic ethnographic studies played an important role in advancing our understanding of intercultural variation in the sources and expression of psychological distress. For example, the works of Arthur Kleinman (e.g. Kleinman Reference Kleinman1980) have become foundational texts in the now well-established field of cultural psychiatry, from which we derive commonsensical aspects of contemporary clinical care, such as ‘cultural idioms of distress’ and ‘cultural competence’.

Ethnography’s commitment not only to investigating culture but also to interrogating its own theoretical nomenclature has led to critiques of the way culture is envisaged in clinical models of cultural competence. For example, Santiago-Irizarry (Reference Santiago-Irizarry2001), ethnographically investigating a culturally sensitive mental health programme designed for Hispanic people in the USA, found that clinicians presented culture as static, conflated culture with race and ethnicity, and inadvertently placed blame on patients’ cultures. As Kleinman & Benson (Reference Kleinman and Benson2006) observed, ‘Culture is often made synonymous with ethnicity, nationality, and language. For example, patients of a certain ethnicity – such as, the “Mexican patient” – are assumed to have a core set of beliefs about illness owing to fixed ethnic traits’. Other anthropologists have added that cultural models do not acknowledge diversity within groups, fail to recognise biomedicine as a cultural system itself and often obscure structural power imbalances by emphasising cultural differences (Carpenter-Song Reference Carpenter-Song, Schwallie and Longhofer2007).

This ethnographic evidence has led Metzl & Hansen (Reference Metzl and Hansen2018) (see also Aggarwal Reference Aggarwal2025) to argue that psychiatrists ought not just to be culturally competent, but also structurally competent. By this they mean not just appreciating and effectively interacting with individuals from diverse cultural backgrounds, but also bringing into their clinical practice an understanding of how societal structures, policies and power dynamics affect mental health and healthcare access.

Ethnography’s unique strength – its interpretation of social action in context – is also its weakness. It is difficult to generalise from ethnographic findings because they are locally and historically specific. Ethnography cannot illustrate universal laws or broadly applicable forces, but it can illuminate how they are experienced by particular people at particular times and why. Ethnography focuses on understanding rather than explanation and prediction.

Coming back to the Francis and Shanley reports, both focus on negative cultures. Culture is depicted as a monolithic organisational property, manifesting in clinical practice, for which its leaders must be held accountable. Ethnographic research depicts culture differently, as a site of conflict and flux shaping and shaped by institutional, structural and political constraints. Patients and staff manoeuvre these constraints as best they can, using the resources they have. An ethnographic sensibility would allow investigators not just to impute the existence of negative cultures, but to stand back and interrogate how the particular beliefs, practices and behaviours of managers and clinicians were able to crystallise in the ways that they did. It is precisely this level of minute consideration that may help to advance improvements in clinical care.