The Oxford Research Encyclopedia of Education is now available via subscription. Visit About to learn more, meet the editorial board, or learn about the subscriber services.

Show Summary Details

Page of

 PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, EDUCATION ( (c) Oxford University Press USA, 2018. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 18 December 2018

Interviews and Interviewing in the Ethnography of Education

Summary and Keywords

Interviews are frequently used in ethnographic research, but it is argued that they pose particular difficulties in interpretation. While ethnographers are interested in understanding how people construct and interpret cultures in their natural settings, interviews are based on rules that counteract most normal interactions. Thus interviews in ethnography can only be interpreted within the context of that wider ethnography and the data generated has to be tested against other data generated by different means and data generated in other interviews.

Although some ethnographers avoid the use of interviews, others use a range of different forms of interviews. It is argued that Basil Bernstein’s concepts of classification and framing can be used to clarify the range of forms and to highlight the potential relationships between the form of interview and class, gender, and ethnicity.

Keywords: ethnography, interviews, Basil Bernstein, classification, framing


We live in a world where interviews are pervasive. We are stopped in the street by people who claim to want to interview us, but who are often trying to sell us something. We watch television interviews where politicians avoid questions and present answers to alternative questions in an attempt to make their policies or activities seem more acceptable. We read magazine articles that claim to show us how to answer possible tricky questions in a job interview so that we can present our “weakest points” as being positive strengths.

Interviews in ethnography are not like any of these popular conceptions of the nature of interviews. First, they are research interviews. The aim of these interviews is not to trick the person being interviewed into saying something he or she may later regret, but to establish some kind of “truth” about his or her experiences, behaviors, interpretations, or views.

Second, with a few possible exceptions in policy research, the people interviewed are not selected because they are well-known or famous but because they belong to a particular group, or hold a particular position within an organization or group. They will usually not be identified by name in any publication related to the research. In ethnographic interviewing, the individuals selected will not be seen as representative of that group, but just as one of that group and thus having some of the possible range of experiences, behaviors, and so on that a person in that group can have.

Third, interviews in ethnography are exactly that—they are interviews that are conducted within a wider ethnography. They are just one of several ways by which ethnographers generate data about the culture of a particular group that is the focus of concern. What a person says in a particular interview will always be interpreted by the researcher taking into account the time, place, and wider situation in which the interview was conducted. What is said will not be taken as the only possible thing that might have been said about an issue and will be tested against other data generated in different ways and against what is said by the same person or others in different interviews.

The Nature of Ethnography

I wish to argue here, as elsewhere (Walford, 2009), that ethnography is not another word for qualitative research. It is a separate research strategy that has developed over the past century and entails far more than simply qualitative research. Indeed, I argue that the distinction between qualitative research and quantitative research is somewhat simplistic and unhelpful when it comes to trying to understand cultures. I advocate what might be called a traditional form of ethnography within educational research. The common idea of thinking about ethnography as being synonymous with qualitative research misrepresents traditional ethnography within education and schooling, as early practitioners used a wide variety of what are now thought of as qualitative and quantitative methods. While ethnographers were unlikely to use sophisticated statistical analysis, they often generated quantitative data as well as qualitative field notes and descriptions. Some of the classic educational ethnographies (such as Hargreaves, 1967; Lacey, 1970; Ball, 1981; or Becker, Geer, Hughes, & Strauss, 1961) present a considerable amount of quantitative data to support their arguments.

Quantitative claims, which are frequently made in ethnographies (if only in terms of “many” or “most”), require quantitative data, so the use of structured observation, time-sampling, and even surveys may be required in addition to more open-ended participant observation and more semi-structured interviewing. The methods used depend upon the research questions that the study eventually tries to answer. Thus, one very noticeable feature of early sociology of education such as Ball’s (1981) study of a comprehensive school is the diversity of ways of generating data that were used. Observations were made in a multitude of contexts: in classrooms, while accompanying groups on school visits, during invigilation of examinations, while playing cricket, in the wider community, and so on. Semi-structured interviews were conducted with pupils and teachers, questionnaires were circulated (including socio-metric questionnaires), pupil diaries were kept, school records and registers were examined. The results of the research were presented in a similar variety of ways: with figures, diagrams, and charts alongside quotations from interviews and observed naturally occurring conversations.

Moreover, for a piece of research to count as ethnographic, it must rely on more than just interviews—interviews are only one of the research tools that researchers may use within an ethnography, for the very interpretation of interviews demands that other methods of generating data are available to test the interview data against. This is not to say that researchers cannot use interviews alone in their research—it is simply that this is not ethnography. There are many research questions that are not centered on culture, and far from all research questions are amenable to ethnography.

It worth trying to define what is meant by ethnography more fully. Hammersley and Atkinson (2007, p. 3), for example, start their discussion in the following way:

In terms of data collection, ethnography usually involves researcher participating, overtly or covertly, in people’s daily lives for an extended period of time, watching what happens, listening to what is said, and/or asking questions through informal and formal interviews, collecting documents and artefacts—in fact, gathering whatever data are available to throw light on the issues that are the emerging focus of inquiry.

Alternatively, some writers favor lists. One example, with which I was involved, was put forward by the editors of Ethnography and Education, who listed in the first issue what they saw as the seven main features of ethnography:

The key elements of ethnographic research applied to the study of education contexts are:

  • the focus on the study of cultural formation and maintenance;

  • the use of multiple methods and thus the generation of rich and diverse forms of data;

  • the direct involvement and long-term engagement of the researcher(s);

  • the recognition that the researcher is the main research instrument;

  • the high status given to the accounts of participants’ perspectives and understandings;

  • the engagement in a spiral of data collection, hypothesis building and theory testing—leading to further data collection; and

  • the focus on a particular case in depth, but providing the basis for theoretical generalization.

  • (Troman, Gordon, Jeffrey, & Walford, 2006, p. 1).

While there are some differences between the description and list provided, these writers claim that there is a set of specific criteria that have to be met before a study can be considered to be ethnographic. There needs to be long-term engagement, the use of multiple research methods, and the generation of rich data. The research process also needs to be theory-led and systematic. Understanding is not achieved through chaotic or biased processes, but by systematic and well-ordered generation of data appropriate to the task.

The Nature of Interviews

Ethnography focuses on research questions that are concerned with the complexity of trying to understand cultures. The fundamental question to be answered is “How do they do things around here?” (Deal, 1985). This is a very different type of question from that which can be answered through survey or experimental research, and ethnographic research grew, in part, as a reaction to the positivistic and experimental research that once held sway. Survey research through questionnaires or interviews with representative samples might be seen as reliable and the results generalizable to a wider population, but they could generate answers only to simple research questions. Experimental methods were castigated as setting up unreal situations such that the results could not be expected to be valid or generalizable to wider circumstances. In contrast, ethnography, with its long-term involvement and desire to understand activities in their natural settings, was thought to bring greater validity, as the everyday activities being investigated would be disturbed as little as possible.

Schools and school classrooms, for example, provide a physical space in which teachers and students interact according to rules that are co-constructed. But the culture created is complex and understood somewhat differently by each of the participants. No matter how hard the teacher may try, the shared culture is not made explicit, and participants may not even be able to recognize or articulate some of its aspects. For example, some early studies of classrooms focused on interactions between participants in terms of gender or ethnicity, and found clear biases in the ways that some children from ethnic minorities were treated, alongside major differences between interactions between teachers and male and female students. But it was not that teachers necessarily deliberately treated students differently according to gender or ethnicity and, when confronted with evidence, they were often shocked by the findings. Ethnographers recognized that interviews with teachers would not have been able to uncover the disjuncture between what people think they do and what they actually do, and that observation was a fundamental necessity in generating particular data on cultures.

It is worth remembering what a strange situation an interview is. It cannot be defined as “a conversation with a purpose” as some commentators suggest—all conversations have purposes, if only to pass time amiably. In contrast, in a research interview in its most structured form, the socially accepted rules of conversation and reciprocity between people are suspended. One person takes the lead and asks a series of questions of the other. The other has agreed that this is to be a special form of verbal interaction and is prepared for his or her views to be continuously questioned without the usual ability to be able to return the question. There is no natural “turn-taking” and ability for both to control the flow of topics. Instead, the topics to be covered are under the control of the “interviewer,” and the “interviewee” is expected to have opinions or information about each of the questions asked. Moreover, and most strangely, what the interviewee says in this particularly odd situation is usually taken to have lasting importance—it is recorded for future analysis. This is not a transitory conversation but one that is invested with future significance. Given that the interview is such an extraordinary situation, it is remarkable that so many researchers take the data so generated as reliable and valid.

I have argued elsewhere (Walford, 2017) that concern over the validity of interviews is far from new. How interviewees respond to questions may well depend upon the time of day or year, the weather, and external events. Crucially, the appearance, gender, ethnicity, clothing, accent, and other variables associated with the interviewer will also influence what is said.

We know that interviewers and interviewees co-construct the interview and that the replies to questions are produced for that particular occasion and circumstance. Interviewers and interviewees take part in a performative and rhetorical interchange and will select their words with care and will moderate what they have to say to the particular circumstances. If we put to one side the epistemological question of whether or not there is any ultimate “reality” to be communicated, the interviewees may have incomplete knowledge or faulty memory. They will always have subjective perceptions that will be related to their own past experiences and current conditions. At best, interviewees will give only what they are prepared to reveal about their subjective perceptions of events and opinions. These perceptions and opinions will change over time, and according to circumstance. They may be at some considerable distance from any “reality” as others might see it.

Some critics of interviews go even further and propose that researchers need to move beyond the traditional cooperative paradigm and recognize the underlying conflictual nature of society. In his typical colorful language Douglas (1976) argues that:

In its most extreme form [. . .] the cooperative paradigm of society assumed it is possible to ask members what is going on and they will tell. Yet everyone knows when he [sic] thinks about it that only the naive, the innocent, the dupe takes this position all the time in everyday life. Rather, all competent adults are assumed to know that there are at least four major problems lying in the way of getting at social reality by asking people what is going on and that these problems must be dealt with if one is to avoid being taken in, duped, deceived, used, put on, fooled, suckered, made a patsy, left holding the bag, fronted and so on.

(Douglas, 1976, p. 57)

Douglas argues that researchers should assume that people and groups are in conflict with one another, and that their aims and objectives often clash. He describes the problems that we encounter in interviews in terms of misinformation, evasion, lies, and fronts, and follows this with a detailed exposition of the problems of taken-for-granted meanings, problematic meanings, and self-deception. In places, ethical questions are not always at the forefront of his suggested strategies, but the overall point is well made: what people tell us in interviews is often not to be taken at face value. He argues that this is particularly true when the interviewer asks about aspects of experience that are of special importance to the person being interviewed—for example, issues of sex, money, and power.

But misinformation, evasion, lies, and fronts are only part of the problem. These are the more explicit and conscious ways in which researchers may be misled. There are also taken-for-granted meanings, problematic meanings, and self-deception. We do not have full control over our lives—they are not inherently coherent. They are result of chance and circumstance as much as our own activities and plans. The unexpected happens; the expected does not. We act as if we will live forever (or, at least, for a long time yet), but we may be dead tomorrow. In this uncertainty we all try to make sense of our own worlds, and the interview is one occasion when we try to do so in a semi-public forum. We try to present a reasonably rational image of our own uncertainty. People construct accounts about themselves, their activities, and beliefs that are far more coherent than the lived reality.

Yet, in interviews, and in particular in the transcribed versions of interviews, a coherent and constructed account gives the impression of permanence to something that is inherently transitory. It becomes a text that can be edited, copied, and recontextualized. The original event becomes data to be entered into a qualitative analysis package, and segments become examples drawn from the data bank at the touch of a key. In short, the transcribed interview encourages the possibility of the spoken word being taken too seriously. The phrase that someone happened to have used on a hot Monday afternoon following a double mathematics class gets wrenched out of its context and presented as if it represented the “truth” about one person’s views or understandings.

Observation Within Ethnography

Participant observation is the key method that is used in ethnography—where the researcher takes on a role within the research setting and observes what is going on. In the 1970s and 1980s, before ethics committees made it almost impossible to conduct covert research, there were numerous studies of learning in sites other than schools. One of the best known of these is Patrick’s (1973) study of a Glasgow gang, where he briefly took on the role of a peripheral member of one of the gangs. Edgerton (1979) provided an account of activity on an urban beach and shows how the users of the beach learn to act such that social order is constructed. The first part of the notorious book by Humphreys (1970) focuses on male public toilets as a learning location—here men learn how to engage in sexual activity with other men without any verbal communication between them. All of these studies, and more, were conducted covertly and thus no interviews were used. Instead, participant observation was the core of the fieldwork.

Even where research is openly conducted some researchers have found that they do not need to conduct interviews to answer their research questions. A recent study by Clerke and Hopwood (2014; Hopwood, 2016) researched professional workplace practices and learning within a residential unit in Sydney where each week up to 10 troubled families with young children were equipped with a range of strategies to help them respond to their children and eventually bring about positive changes in family life. In this fieldwork researchers observed, took photographs, collected documents, recorded naturally occurring verbal interactions, and even videoed some activities, but identifiable interviews played no part. Interviews were not seen as being able to help answer the research questions.

Similarly, Barley (2014) used an array of methods in her study of identity and social interaction in a multiethnic classroom of early years’ children. This included a great deal of observation (which, as Delamont [2014, pp. 146–163] reminds us, includes smell, touch, hearing, and tasting, as well as looking), and also various activities with the children to generate data within the classroom environment. Again, interviews played little part.

So Why Use Interviews at All?

In spite of the many problems previously outlined, many ethnographers still use interviews within ethnography, and there are a variety of reasons why they do so.

One reason is simply that other people think that interviews are important. For example, those who allow initial access to the research site often feel that interviews should be part of the research. Before I was granted access to one of the major British private schools to conduct a term-long ethnographic study, the head teacher insisted that I conduct interviews with a range of staff and students. His concern was with size and representativeness, as he wanted to ensure that I obtained information from a wide range of people and did not focus on just a small group, but I was certainly glad that I agreed to his demand. I found that the head teacher was far from being alone in seeing interviews as central to research. It became clear that many of the teachers also saw it as important and were reluctant to allow me to observe until they had had their chance to be interviewed and put forward their views. Thus I found that the teaching staff and students wanted to be interviewed, and this then facilitated access to classrooms and other places for observation and other forms of data generation.

Interviewing at the start of an ethnographic research period does, however, have disadvantages. Ideally, interviews might better be conducted some way into the project when relationships have been established and those involved are comfortable in their knowledge of the researcher and what he or she is aiming to achieve. Interviewing early may help with access, but also may lead to more evasive and misleading responses to questions. With early interviews there has not been enough time for trust to have been developed, so that there is really a need for multiple interviews if these interviews are to be used as a source of factual data. Comparing what is said in early and later interviews can be revealing.

Beyond the strategic, there are several other reasons why ethnographers might use interviews in their studies. For example, it is often important to know something of the informant’s history in interpreting what is seen in the classroom or other learning sites, and that is most conveniently obtained directly through an interview. Similarly, interviews can be used to try to understand what is going on during a period of observation, or to get various participants’ views after an event. Interviews can give information on activities that occurred outside of the period of study, or during the period of study but when the research did not obtain access. In all cases the key is that interviews need to be seen as being potentially flawed, and they can be interpreted only through integration and challenge with other data generated in the ethnographic research process. Interviews in ethnography must be seen as a particular form of observation of a situation in which the interviewer has set up a strange theatrical performance where both participants perform only partially scripted roles.

What Type of Interviews May Be Used?

Many commentators go to considerable lengths to distinguish between such descriptors as structured, semi-structured, loosely structured, and unstructured interviews. I will not do so here, as a single interview can contain elements of all of these. At some points in an interview, explicit questions might be asked that require a yes or no answer, or a precise number. At other times the ethnographer might ask the interviewee to describe an activity that has just occurred, or ask the person to describe some of the good or bad parts of their teaching experiences. It is likely that most interviews within ethnographies will use many open questions where the interviewees are able to put forward their own views—or, at least, those views packaged for the particular circumstances of the interview. These open questions may be followed by more probing questions that respond to what has been said and ask for more details or further explanations.

Some time ago I conducted some research where I interviewed a small sample of well-known ethnographers of education (Walford, 2007). The main suggestion that I put forward then was that it was fruitful to use Basil Bernstein’s ideas of classification and framing to describe and think about the nature and process of ethnographic interviewing. These two concepts were introduced in a much-quoted article (Bernstein, 1971), where they relate to the form and structure of the curriculum and the transmission of educational knowledge. Two educational knowledge codes are discussed, differing according to the underlying principles that shape curriculum, pedagogy, and evaluation. A curriculum of a collection type is characterized by strongly bounded knowledge areas with little linkage between them. On the other hand, an integrated curriculum emphasizes the interdependence of various area of knowledge and attempts to transcend traditional boundaries. Bernstein argues that any structure for the transmission of knowledge will symbolically reproduce the distribution of power in society, and introduces the concept of classification to clarify this relationship:

Where classification is strong, contents are well insulated from each other by strong boundaries. Where classification is weak, there is reduced insulation between contents, for the boundaries between contents are weak or blurred. Classification thus refers to the degree of boundary maintenance between contents.

(Bernstein, 1971, p. 49) (original emphasis)

The concept of frame refers to the strength of the boundary between what may be transmitted and what may not be transmitted in the pedagogic relationship. It indicates “the degree of control teacher and pupil possess over the selection, organization, pacing and timing of knowledge transmitted and received in the pedagogical relationship” (Bernstein, 1971, p. 50). The strength of framing thus refers to the range of options available to teacher and taught in the control of what is transmitted and received.

Bernstein’s work has been applied to very many different empirical situations, and many researchers have found the general framework useful in attempting to clarify a range of educational problems. Some early examples include Walker (1983), who used Bernstein’s concepts of classification and framing within a historical analysis of different social regimes in colleges of education over a century, while Aggleton and Whitty (1985) applied them to a study of the subcultural practices of a group of new middle-class students in an English college of further education. Examples of broader and more flexible usage include Rodger’s (1985) study of a large public inquiry and Walford’s (1981) account of problems within postgraduate research and of the curriculum in major private boarding schools (Walford, 1986).

Applying the concepts of classification and framing to interviewing within ethnography requires a little stretching, but knowledge is still being transmitted from one person to another in a specific situation. The main difference is that, for the most part, knowledge is actually being sought by the interviewer rather than information being given by a teacher. The power within the relationship is thus rather different in that, while the interviewer generally has greater power to classify and frame the situation, it is the interviewee who can (re)construct the information transmitted, and who has the ultimate sanction of withholding information.

Classification and Framing of Interviewing

Classification is thus concerned with separation. In a similar way to there being a collection or integrated code in terms of the curriculum, we can think of there being a collection or integrated code with regard to the method and nature of generating data through interviews in ethnography. There is a range of ways in which interviews are conducted such that, while some are kept deliberately separate from other activities, others hardy appear to be interviews at all, and may even not be recognized as such by participants. At various points in their studies, ethnographers may sometimes use strongly classified interviews where two people face one another over a desk and recording instrument, and at other times simply discuss events together, with the interviewer later making notes.

In Bernstein’s usage the concept of frame refers to the strength of the boundary between what may be transmitted and what may not be transmitted in the pedagogic relationship. It indicates “the degree of control teacher and pupil possess over the selection, organization, pacing and timing of knowledge transmitted and received in the pedagogical relationship” (Bernstein, 1971, p. 50). By analogy, in interviewing framing refers to the degree of control that the interviewer and interviewee possess over the selection, organization, pacing, and timing of the knowledge transmitted within the interview situation.

The variability that is possible in the nature of interviews can be seen through some quotations from my interviews with ethnographers. For example, in discussing his work with Mexican children, in his interview with me, Bradley Levinson described his early stage of participant observation:

B So a lot of it was really just roaming out and observing early on especially before I got to know . . . but actually the kids didn’t let me do too much of that because a lot of them very early on took an interest in me or a liking to me and they’d see me walking round and they’d go, tsh tsh, come over here and they’d call me over and they’d want to ask me who I was and what I was doing or they’d want to ask me, what did I think about what I’d just observed in class [oh yes] or they just wanted to kind of incorporate me into their friendship dynamics in ways I could bring a kind of prestige to some of these groups by association.

Here we see an example of a form of interview that is hardly distinguished as such. There is weak classification between this activity and other activities—observation, in the form of listening and interacting with the students, shades into interview. From the point of view of the researcher, framing is also weak, for the students were asking questions of the researcher as well as the other way round.

This form of interaction/interviewing is far from unusual. Sara Delamont’s discussion of what she did in her capoeira research (Stephens & Delamont, 2006; Delamont, Stephens, & Campos, 2017) reveals similar weakly classified and framed interviews.

S For instance there is one woman who is basically a dancer who is fascinated by capoeira and she actually borrows books off me and, but she’s equally likely to ask me if I’ve got a book about Alexander the Great or a book about the whirling dervishes, because she’s always making up dance programmes with all sorts of cultural [. . .] and she talks to me about human movement theory and that sort of thing. But I think for her, because she’s not in higher education I think actually I’m her free library. I mean she’s delightful [G yeah, yeah] I like her, but I think she talks to me about things like that because she knows that if she says to me ‘you don’t happen to have a book about witchcraft beliefs in . . . and if so can I borrow it?’ um or ‘do you know anything about Santeria in Cuba and do you happen to have a book about Santeria drumming I could borrow?’

Here, not only is information being generated through everyday conversation, it is again a two-way process with Sara being asked questions as well as her asking them.

Yet both of these ethnographers also conducted interviews that were more tightly classified and framed. In the same capoeira research, Sara Delamont almost defines “an interview” by where it is conducted. I have left some of the indications of hesitation in this transcript as they indicate some of the thinking through of the question of place that happened during the interview.

S oh I do all the interviews here [in her office at the university]

G right, why do you do that?

S oh, because it’s an appropriate space [G ok] and I feel very strongly about that I do all the interviews here

G tell me about what an appropriate space is

S well where else would I do them? [G yeah I agree that’s obviously a problem] um I think there are three reasons why I do them here. Partly because I think it’s, I could do them at the house and that would be fine and that isn’t a problem. I’m not frightened of having these young men or the women, I wouldn’t be frightened of having them in the house or anything like—we could sit and do interviews on my dining room table or something [G yes]. But I’ve gone in for doing it here because I think that signals to people that it is part of a university job [G yes] and when Ben gossips afterwards, I mean Ben will come in and say—the first time he came in here it was kind of ‘hell’ you know ‘what a lot of books.’ But as far as, I said, if I’m interviewing formally about capoeira he comes here and I would expect. I think it’s three things. I think one is it’s to prove that I really am a person in a university because how can they know? [G no, quite, no I agree] and secondly I think it’s because for me it marks it as work [G yeah] and thirdly I think, I think I feel it’s sort of more appropriate um, I wouldn’t, I mean if somebody said come to my flat and interview me there I’d say no, I think I’d say I’d rather you came to me if you don’t mind. Maybe I would go but. . . . When I did the interview part of the St Luke’s study [of a girls’ private school] I actually interviewed them all off school premises [G did you] in my flat [G did you] as a sort of, yeah, I sort of interviewed them off premises in their ordinary clothes and I organised for them to come, they came in sort of friendship groups and sat around and drank coffee and I hooked them out of the friendship group and took them to another room and interviewed them [G right, right interesting] um and I just thought that was right then.

We have already seen that Bradley Levinson generated some data using very weakly classified and framed interviews, but the following quote illustrates that he also used slightly more strongly framed interviews in asking specific follow-up questions as well as strongly classified one-to-one interviews. He indicates that, in this case, the latter were not as productive as he had hoped.

B The interviews ranged from following up on a particular event that I had witnessed in their classroom [G oh really?] so it would be, you know, sometimes there were these key incidents you might say in the fieldwork when there was a big argument that had run between a whole class and a teacher and they were very illuminating of certain, we anthropologists always love these juicy moments of confrontation because of the sort of the fissures that they reveal. So I would try to get perspectives of different groups of students on it, if I noticed what seemed to be some kind of tension in the group, social tension between different friendship cliques within the group I’d want to go round and get their different perspective on it so it wasn’t really any sort of intrinsic part of what I’d call my research design initially. My research design was sort of doing focal student interviews, choosing my focal students and doing a series of so-called talking diary interviews with them, ideally at least three over the course of the school year. I would say as an aside that I think I stayed too wedded to that design even when it became clear to me relatively early on that wasn’t going to be terrible fruitful especially because some of the students were not very good interviewees, they were just too immature especially the boys they were just too immature or too uncomfortable in a one-on-one situation to really articulate things very well.

Clearly there is considerable variability in the framing of interviews. At one end of the spectrum within ethnographic interviewing might be the style used by Lois Weis when conducting her re-interviews more than a decade later of the young people who were part of the Working Class Without Work project (Weis, 1995, 2004). She explained her way of working at this point as:

L When I re-interviewed the Class Reunion kids, I had a page of questions. It was more than a page, two pages of questions. And I actually showed them the questions I had intended in order to get their consent. I said “these are the questions that we will be talking around, do you have any objection, here’s the tape recorder button if you ever feel you want to turn it off” and so that is my way of also letting them know we were going to do and getting their permission to tape.

G So in some ways it was more of a tactic to gain access [L: yes] than what you would necessarily follow. Obviously you would follow in general those questions, but

L It’s both. I did follow the questions. Sometimes the questions worked a little less well, but I knew what I was doing then. They were issues I really wanted to cover and they were issues I wanted to cover because I knew they were important by all the work I had done and from books that I had read.

Here Lois Weis is trying to use the interviews to cover particular aspects that fit with her empirical and theoretical concerns and is using a list of questions to ensure strong framing of the situation. The following extract from our interview shows both that strong framing has to be fought for, and that at the end of these same interviews Lois allowed a section that was much more weakly framed when she passed back to them the transcripts of their previous childhood interviews.

G Ok you gradually work your way through the questions and, as I’m doing, presumably you darted around a bit?

L Absolutely, I always dart, yeah. And I feed off what they say. I mean, like you, you know you get pretty good at this after a while and I feed off what they say. So I move with them.

G How do you end the interview?

L The Class Reunion ones were really fun because I gave them a typed version of their transcript from when they were sixteen and they flipped, they just freaked out.

G You gave that after you’d interviewed them did you?

L Yes I gave it to them afterwards. And then we talked about it a little bit, it was more a sort of my closing.

G With the tape recorder still on?

L Yes. It elicited a little less than I thought it would, with the tape recorder on because I think they didn’t think about it.

In contrast, the interviews conducted by Paul Connolly (1998, 2004) with his primary school children were much more weakly framed.

P. I asked a child to nominate two others and I found that a group of three was probably the optimal number for me for that age [G: ok]. With four you started to create dynamics where you could have two against two [G: yeah] and with just two, you know, if one wasn't speaking much it became quite stunted. Three for most of the time actually worked pretty well. So you’ve got a friendship group of three, basically sat down, and there’d be opening questions that I had—just about what were you doing today in the playground or I’d have more general questions about what are you going to do tonight when you get home, or what do you like to do at home, what do you like to play, what do you watch on TV. Just to get the discussion going [G: yeah]. And then really it would take on a logic of its own. So for the most part I really just let the children talk in whatever ways they wanted to. And I had that luxury of being a PhD student where I could spend three days a week for a year in the school so I could take my time [G: sure] and just allow them to discuss whatever they wanted to. And invariably, you know, it’s an inner city school in a multi-racial area, and race came into their discussions very quickly.

Here the circumstances were designed such that the children could take a large degree of control over the framing of the situation. It was deemed appropriate because the research centered on the understandings that these children gave to race and ethnicity. Allowing them considerable power to frame the situation meant that they could exhibit their own worldviews.

At first sight, it might be thought that classification and framing when applied to interviews are overlapping variables rather than being independent—but, while strong framing and strong classification often seem to go together, as do weak classification and weak framing, there are many situations where the strength of each variable is in opposition.

It is far from unusual, for example, for researchers to have to struggle if they wish to impose strong framing. Jan Nespor describes an occasion where he was invited to interview a group of people in the school committee and the framing was fought over:

J. I actually served on, like there was a site based school committee there, site based management committee, that I was a member of for a couple of years and I know one of the members of that group from the church was on it. Some way or another I hooked up with them and I went over and interviewed about ten of them simultaneously. Which really wasn’t my idea but that was like, I said ‘Can I come and interview you about what you’re doing after school and what the school was like in the past?’ and she says, ‘Oh yes we’d love, we’ll bring a whole bunch of people who were there.’ You know, they kind of introduced themselves, and you’d try to take notes so you can reconstruct who said what because everybody sounds the same on the tape. [. . .] And of course with ten people that very quickly became them talking to each other more than me asking questions.

Or there might be particular people who are reluctant to go along with the strong framing that the interviewer might desire. Jan Nespor recounted one such woman:

J. She was very involved in opposition to the school and she would never sit down and be interviewed. She was one of those people who would grab you in the parking lot and tell you everything that had happened to her and talk for twenty or thirty minutes but just didn’t want to have it on tape. And I’d say can I write about that? I mean I didn’t actually because I could never remember. It was like, you know, it’d be the end of the day and you’re tired, and this person just downloads you know tries to download all of this complex stuff and of course by the time I get to the typewriter I can’t remember half of it.

Researching the powerful in education can also challenge both classification and framing. In an edited collection on the topic of researching the powerful (Walford, 1994) there are numerous examples where policymakers were interviewed to try to gain insights into their own role in particular policy decisions. Stephen Ball gives an example of his interview with Sir Keith Joseph where the interviewer and interviewee have somewhat different agendas and struggle to control the framing of the interview. The classification is very strong, but the framing, from the researchers’ point of view, keeps slipping away. Similarly, John Fitz and David Halpin describe an interview with Kenneth Baker where official advisers sat in on the interview and “the interview that followed was dominated by Baker”(Fitz & Halpin, 1994, p. 44). Those who are powerful and used to being interviewed on-record find it easy to make sure that interviews are actually weakly framed even where the interviewer plans it to be strong.


I have argued that there are many problems with interviewing in educational research, but that this does not necessarily mean that interviews should be abandoned within ethnography. Interviews within ethnography should always be recognized as being part of the wider ethnography and must be interpreted in conjunction with data obtained through other research methods or through other interviews. Interviews can inform us of what the person interviewed is prepared to say about a topic in the social context, time, and place of that particular interview. We need to recognize that what is said will be co-constructed in that interview, and will be limited by perception, memory, evasions, self-deception, and more on the part of both interviewee and interviewer, and researchers certainly cannot uncritically take what is said in any particular interview as “the truth” for all times. Interviews have to be treated as generated accounts and performances (Atkinson, 2015, p. 96) and to be recognized as occasions when those interviewed will construct themselves as particular types of people.

This means that ethnographers need to analyze interviews with much more care than is often given to them. We need to focus on what is not said as much as what is said. We need to be cautious in interpreting the words produced in interviews, and try to generate further data about the same topics in a variety of ways and in different circumstances. This means that we need to search for misinformation, evasion, lies, and fronts, and also to look for taken-for-granted meanings, problematic meanings, and self-deception. We need to interrogate the interviews and compare what is said in one account with what is indicated by another interview or research method.

I have used Bernstein’s categories of classification and framing and applied them to the variety of interview situations within ethnography. As with the original formulation of educational knowledge codes, there is a tendency for high classification to be associated with high framing and low classification to be associated with low framing, but two distinct dimensions can be identified.

The more challenging implication of using the framework of classification and framing is that, in Bernstein’s original formulation, the educational success of children of differing classes (and, one might add, genders and ethnicities) will be related to the classification and framing to be found in teaching situations. “Where framing is strong [. . .] social class may play a crucial role [. . .] it often means that the images, voices and practices the school reflects make it difficult for children of marginalized classes to recognize themselves in the school” (Bernstein, 2000, p. 14). One would expect there to be similar relationships between class and other elements of marginalization in the interview process. One advantage of the use of classification and framing as a way of thinking about interviewing is that it may make researchers examine the possible differential validities and reliabilities that various approaches may have with interviewees of different social classes, genders, and ethnicities. It is an empirical question as to which type of interviewing technique is likely to generate the most valid data, and this requires a separate research investigation, but it is inevitable that this relationship between validity and method will be mediated by class, race, and gender.

Author Note

As with much of my writing, this contribution draws upon and develops some previous work, in particular, Walford, 2001, 2007, 2008, 2009, and 2017. This article largely takes a British and European viewpoint. It does so because, sadly, since the late 20th century, in the United States in particular, the two words “qualitative” and “ethnography” have become almost synonymous. Just as seriously, many in the United States use the word “ethnography” to describe studies that are based on interviews only. For example, the Ethnography in Education Research Forum held each year in Philadelphia advertises itself as the “the largest annual meeting of qualitative researchers” and has papers using a wide range of qualitative methods, but disappointingly few that use ethnography. The U.S.–based journal Anthropology and Education Quarterly has many articles that are based on interviews only rather than the traditional focus of anthropology on ethnography. But in the United Kingdom and most of Europe the distinctions remains clearer, with, for example, the European-based journal Ethnography and Education only publishing articles based on the traditional view of ethnography and the annual Ethnography and Education Conference held at Oxford not accepting interview-only papers.

Further Reading

Beach, D., Bagley, C., & Maques da Silva, S. (Eds.). (2018). The Wiley handbook of ethnography of education. London: Wiley.Find this resource:


    Aggleton, P. J., & Whitty, G. (1985). Rebels without a cause? Socialization and subcultural style among the children of the new middle class. Sociology of Education, 58, 60–72.Find this resource:

      Atkinson, P. (2006). Rescuing autoethnography. Journal of Contemporary Ethnography, 35(4), 400–404.Find this resource:

        Atkinson, P. (2015). For ethnography. London: SAGE.Find this resource:

          Ball, S. J. (1981). Beachside comprehensive: A case study of secondary schooling. Cambridge, UK: Cambridge University Press.Find this resource:

            Barley, R. (2014). Identity and social interaction in a multi-ethnic classroom. London: Tufnell.Find this resource:

              Becker, H. S., Geer, B., Hughes, E. C., & Strauss, A. L. (1961). Boys in white: Student culture in a medical school. Chicago: University of Chicago Press.Find this resource:

                Bernstein, B. (1971). On the classification and framing of educational knowledge. In M. F. D. Young (Ed.), Knowledge and control (pp. 47–69). London: Collier-Macmillan.Find this resource:

                  Bernstein, B. (2000). Pedagogy, symbolic control and identity (rev. ed.). Boulder, CO: Rowman & Littlefield.Find this resource:

                    Clerke, T., & Hopwood, N. (2014). Doing ethnography in teams: A case study of asymmetries in collaborative research. London: Springer.Find this resource:

                      Connolly, P. (1998). Racism, gender identity and young children. London: Routledge.Find this resource:

                        Connolly, P. (2004). Boys and schooling in the early years. London: RoutledgeFalmer.Find this resource:

                          Deal T. E. (1985). The symbolism of effective schools. Elementary School Journal, 85(5), 601–620.Find this resource:

                            Delamont, S. (2005). No place for women among them? Reflections on the axé of fieldwork. Sport, Education and Society, 10(3), 305–320.Find this resource:

                              Delamont, S. (2014). Key Themes in the Ethnography of Education. Los Angeles: SAGE.Find this resource:

                                Delamont, S., Stephens, N., & Campos, C. (2017). Embodying Brazil: An ethnography of diasporic capoeira. London: Routledge.Find this resource:

                                  Douglas, J. D. (1976). Investigative social research: Individual and team field research. London: SAGE.Find this resource:

                                    Edgerton, R. B. (1979). Alone together: Social order on an urban beach. Berkeley: University of California Press.Find this resource:

                                      Fitz, J., & Halpin, D. (1994). Ministers and mandarins: educational research in elite settings. In G. Walford (Ed.), Researching the Powerful in Education. London: UCL Press.Find this resource:

                                        Hammersley, M., & Atkinson, P. (2007). Ethnography: Principles in practice (3rd ed.). London: Routledge.Find this resource:

                                          Hargreaves, D. (1967). Social relations in the secondary school. London: Routledge and Kegan Paul.Find this resource:

                                            Hopwood, N. (2016). Professional practice and learning: Times, spaces, bodies, things. London: Springer.Find this resource:

                                              Humphreys, L. (1970). Tearoom trade. London: Duckworth.Find this resource:

                                                Lacey, C. (1970). Hightown Grammar. Manchester, UK: Manchester University Press.Find this resource:

                                                  Nespor, J. (1997). Tangled up in school: Politics, space, bodies, and signs in the educational process. Mahwah, NJ: Lawrence Erlbaum.Find this resource:

                                                    Patrick, J. (1973). A Glasgow gang observed. London: Eyre Methuen.Find this resource:

                                                      Rodger, J. J. (1985). Natural justice and the big public inquiry: A sociological perspective. Sociological Review, 33, 409–429.Find this resource:

                                                        Stephens, N., & Delamont, S. (2006). Balancing the Berimbau: Embodied ethnographic understanding. Qualitative Enquiry, 12(2), 316–339.Find this resource:

                                                          Troman, G., Gordon, T., Jeffrey, B., & Walford, G. (2006). Editorial. Ethnography and Education, 1(1), 1–2.Find this resource:

                                                            Walford, G. (1981). Classification and framing in postgraduate education. Studies in Higher Education, 6(2), 147–158.Find this resource:

                                                              Walford, G. (1986). Ruling-class classification and framing. British Educational Research Journal, 12(2), 183–195.Find this resource:

                                                                Walford, G. (2001). Doing qualitative educational research: A personal guide to the research process. London: Continuum.Find this resource:

                                                                  Walford, G. (2007). Classification and framing of interviews in ethnographic interviewing. Ethnography and Education, 2(2), 145–157.Find this resource:

                                                                    Walford, G. (2009). For ethnography. Ethnography and Education, 4(3), 273–284.Find this resource:

                                                                      Walford, G. (2017). Recognisable continuity: A defence of multiple methods. In D. Beach, C. Bagley, & S. M. de Silva (Eds.), The Wiley handbook of the ethnography of education. London: Wiley.Find this resource:

                                                                        Walford, G. (Ed.). (1994). Researching the powerful in education. London: Routledge.Find this resource:

                                                                          Walford, G. (Ed.). (2008). How to do educational ethnography. London: Tufnell.Find this resource:

                                                                            Walker, M. (1983). Control and consciousness in the college. British Educational Research Journal, 9, 129–140.Find this resource:

                                                                              Weis, L. (1995). Working class without work: High school students in a de-industrializing economy. London: Routledge.Find this resource:

                                                                                Weis, L. (2004). Class reunion: The remaking of the American white working class. London: Routledge.Find this resource: