From: Banister, P., Burman, E., Parker, I., Taylor, M. & Tindall, C. (1994) Qualitative methods in psychology: A research guide, Buckingham: Open University Press
Conducting interviews is a complex, labour-intensive and uncertain business, fraught with tricky issues that social scientific researchers, and particularly psychologists, are often ill-equipped to address. This is because the emphasis on detachment and the distance structured between researcher and researched within most psychological theory and research instruments is rudely challenged by the face-to-face research interview. However much this is warded off by professional and personal defences, an interview is at some level inevitably a personal and sometimes intimate, as well as public, encounter. This chapter briefly outlines the rationale for doing interviews and different theoretical models of the interviewing process. The focus here is on the researcher's role, and the analysis section will draw on examples of interviewer and interviewee relations to elaborate this further. At the outset I should make it clear that, although guidelines for good practice are presented here, as with some other qualitative research methods, the diverse and specific nature of interviews means that no blueprint of interviewing practice or analysis can be absolutely determined in advance and in abstraction from the topic and context of a particular inquiry. Rather, the aim here is to highlight some of the issues that need to be addressed when designing, conducting and analysing interview-based research. It should also be noted that this chapter, as a reflection of most of the research literature and practice, is concerned with one-to-one, face-to-face interviews. While questions about the conduct of the research and the power of the researcher discussed here also apply to group discussions (which are increasingly gaining recognition as contexts of research, from market research to action research), these may be either magnified or mitigated by distinctive group processes.
While the content of this chapter covers issues sometimes addressed under the titles of 'social research' or 'ethnography', we are dealing here with what might broadly be called 'semi-structured' approaches to interviewing. We will use the term 'thematic analysis' for the process of making sense of the interview. There is a double contrast implied in the account of interviewing practice presented here: with structured approaches, which are usually quantitative and closer to questionnaires in structuring the interviewee's responses, and with so-called unstructured approaches, which for reasons argued below we regard as at best a disingenuous and sometimes a dangerous misnomer for refusing to acknowledge prior expectations or agendas. The position taken here is that assumptions structure all research, and the least we can do is to recognise this and theorise the impact of these assumptions. Better still, we can plan and articulate our starting assumptions so as to scrutinise and promote the research goals. Accordingly, the questions raised in this chapter tie in closely with the discussion in Chapter 8 on feminist research, and some of the more critical issues about interviewing practice are taken up there.
There are four main reasons for conducting interviews. First, uniting the many models of interviewing is a concern with subjective meanings (the meanings the participants accord to the topic of the interview) rather than with eliciting responses within a standard format for comparison with other individuals or groups.
Second, interviews can permit exploration of issues that may be too complex to investigate through quantitative means. That is, given the latter's aim to simplify phenomena, they can misrepresent the nature of the questions under investigation. For example, if you wanted to explore roles, relationships and ethics with a particular professional group, or even to compare perceptions of a service between service providers and recipients, it is unlikely that you would gain a sufficiently sensitive and incisive grasp of your participants' concerns by administering a questionnaire with rating scale categories. This might be not so much because the scale does not address the correct questions (although this might also be true), as because the views of the participants cannot be readily representable within that form. Hence, holding inconsistent, contradictory views is not necessarily a function of faulty reasoning, but rather may be a reflection of the real contradictions and complexities of the way the service works in practice.
Your aim in using a semi-structured interview may be to explore precisely those areas where your interviewee perceives gaps, contradictions and difficulties. Hence another advantage of using a less structured approach is that you can tailor your questions to the position and comments of your interviewee, and you are not bound by the codes of standardisation and replicability to soldier on through your interview schedule irrespective of how appropriate it is for your interviewee.
Within this approach, then, you should respond to and follow up issues raised by your interviewee, including ones that you may not have anticipated. In this sense, semi-structured interviewing, as a more open and flexible research tool, can document perspectives not usually represented (or even envisaged by researchers), and hence the approach can empower disadvantaged groups by validating and publicising their views (e.g. Mishler 1986). While in an experiment the key question is specified in advance as the hypothesis (or so the story goes), and that question is (supposedly) the only question that the experiment addresses, in an interview the focus of the interview can be (although is not necessarily) a matter of negotiation.
Third, doing interviews is a salutary lesson in research involvement and practice. Without the 'safe' distance of a one-way mirror or the position of the detached manipulator of variables, as an interviewer one is forced to confront one's own participation within the research. We can take this point further to reflect on whether this lesson is particular to interviewing, or whether questions of the social construction of research 'data' reverberate further with implications for all empirical work. At any rate, conducting interviews demands consideration of reflexivity in the research process, extending from the devising of the research question, to identifying and setting up interviews with informants, to the interview itself (your role, how you were seen by the interviewee, your reflections on the process), and including the work done to transform an interactive encounter into a piece of written research.
Fourth, necessarily associated with the process of making visible your own work in the construction of your material, is the question of power relations in research. An early account of interviewing (Bingham and Moore 1959) describes the interview as a 'conversation with a purpose'. We should stop to consider whose purposes the conversation is pursuing. Research sets up, and is conducted within, power relationships. We need to attend to these in terms of both the morality-politics of research practice and the academic criteria of adequate evaluation of research (though such a strict separation is of course impossible to maintain once we consider these issues). Some models of research relationships try to do more than acknowledge the structural power relationships set up by research (see the chapters on feminist research and action research), to mitigate, challenge or even reverse traditional power dynamics. The move from designating the people who form the focus of the research as 'subjects' to 'interviewees' or ‘participants' or 'informants' or 'co-researchers' reflects attempts to do research 'with' rather than 'on' people. Of course, the research relationship is only one of various structural power relations that can enter into the research. We should also, therefore, consider the extent to which class, 'race', gender and age relations (for example) interact with the interviewing relationship. Again we can reflect on the extent to which these issues are specific to interviewing, or, although perhaps particularly visible here, are just as relevant to other forms of research practice. Nevertheless, we need to maintain an 'interpretive vigilance' (Figueroa and Lopez, 1991) to ward off the ways researcher control is implicitly structured and exercised within research instruments claiming to be participative and consultative.
Models of interviewing
Broadly speaking, four approaches inform interviewing practice: ethnographic, 'new paradigm', feminist and postmodernist. While these approaches can overlap and combine, each has its own language and way of conceiving the research process and relationship. So while they have much in common with each other, it is worth identifying a few contrasts or points of tension within interviewing style and interpretation here. In all approaches, however, reflexivity is accorded a key role, in the sense of the researcher reflecting on her or his own experience and role within the conduct of the research.
While ethnographic work highlights informants' expertise and the dependence of the researcher on the informant for access to her or his subjective rules, meanings and cultural life, there is a clear role demarcation between researcher and researched in determining the research topic and outcome (although this is changing within contemporary anthropological work: see Nencel and Pels 1991). Further, notwithstanding its ethos of eliciting and representing descriptions, we should not lose sight of how even ethnographic work still requires prior identification and structuring of themes to be investigated. On this, James Spradley (1979: 55) provides a clear account of the differences between an ethnographic interview and an 'ordinary conversation . There are similarities here with Jean Piaget's clinical interview process, where it is argued that 'the good practitioner lets himself [sic] be led, though always in control, and takes account of the whole of the mental context' (Piaget 1929: 19).
In contrast, in 'new paradigm' research (Reason and Rowan 1981), while following the ethos of valuing what people say and treating this as meaningful and informative, research is viewed as a collaborative enterprise which not only involves the full participation of the interviewees but
also incurs responsibility on the part of the researcher to be accountable to, and in some cases to conduct research agendas according to the demands of, the participants (see Chapter 1). Here we see the traditional model of researcher-researched relations undergoing upheaval as the researcher strives to carry out research in a non-exploitative, non-dehumanising way.
Discussions of feminist methodology also take as central issues of power in the conduct of research. But rather than focusing only on the interpersonal relationship set up within the research encounter, feminist approaches attend in addition to wider questions of power as they enter into the funding, popularisation and uses of research (e.g. Spender 1981). Moreover, they often treat power not as something that can be removed from research, but rather as an ever-present dynamic that needs to be acknowledged as structuring the interaction in diverse ways. In this sense feminist analyses of power in terms of the social positions occupied by interviewees, and (re)produced within interviews, go beyond those offered in 'new paradigm' accounts - most noticeably, but not exclusively, in terms of gender.
Finally, there are accounts of research drawing on post-structuralist and postmodernist writings to critique traditional models of research. This might include social constructionist and narrative approaches to research (e.g. Mishler 1986; Steier 1991). Of particular relevance here is the questioning of the presumption that participants within research share the research goals. The changes to which the research is directed may well be worthy, but may be of no immediate benefit to the informant at whose expense careers are gained and whose experience is subordinated to a preconceived or more or less imposed interpretive framework (see Gubrium and Silverman 1989; Opie 1992). Critiques along these lines invite attention to the variety of interpretations that can and will be made by different parties to the research encounter, and therefore also call for a principled scrutiny of our work of interpretation as researchers. In addition, more transformative research practice would seek to identify and address this multiplicity of interpretations in terms of research goals.
Constructing and selecting interview material
It is worth remembering that work done before the actual conduct of the interview is usually amply repaid in terms of its success and ease of analysis. First, you will have arrived at a topic to research, but you should clarify the rationale for doing this. Second, you should specify who would best exemplify the perspectives or range of perspectives relevant to your research question. Third, you should generate an interview schedule. At early stages in the planning of the research this may simply be a list of headings which you can elaborate in more detail once you have sorted out who your participants are, but it is worth doing this work now so that you have a clearer focus for when you approach your participants. Fourth, now that you have decided what kinds of people you want to interview, you need to contact them. It is very important to consider the impact of the route by which you contacted your participants in terms of how this structures the ways they see you, so that, for example, if you are interested in experiences of social or health services delivery it may be difficult to dispel the image of being associated with evaluation or treatment if you initially approach them via a medical or legal agency. It may, however, be impossible to avoid such constraints, but you should at least theorise how this may limit the form and content of the accounts you elicit.
What your prospective interviewees see the study as being about will also be central to their decision about whether to participate, and, in line with codes of practice about 'informed consent', you should be as open as possible about your aims. This may include outlining the kinds of areas or questions you want to discuss with them, and this information can do much to allay participants' anxieties or reservations. You should also at this point discuss what records you want to make of the interview, as in seeking permission to audiotape, for example; it might be helpful to explain why this is useful and how you will use it. Fifth, at this point you should negotiate a research contract with your participant, which includes guarantees of anonymity, a promise to terminate the interview at any point if the interviewee feels uncomfortable, the exclusion from the transcript or other records of anything the interviewee does not wish to be seen by others and a copy of the final report if desired.
While all this may gain you your participants (and if people refuse, you should consider why this may be), you now need to plan the interview itself. First, you will need to elaborate your interview schedule. In qualitative interviews it may not be appropriate to ask your interviewees similar questions; indeed, the 'same' question may have a far from equivalent meaning depending on the interview context, the interviewee's position and the research relationship. Since what you are interested in here is divergence and variety, rather than convergence and replicability, you may be better able to address your general aims by orienting the question to the particular positions of your participants.
Some people like to prepare a detailed interview schedule, with questions addressing all the key issues they want to cover. While this can be reassuring for the researcher, it needs to be treated flexibly in the interview itself since too rigid adherence can intimidate the participant or can fail to follow the participant's train of associations and perspectives. It can therefore be more helpful to have a list of topic areas, with lists of issues you want to cover, arranged so that it is easy for you to check them out in the course of the interview. But in this case the danger is that, while responding to the particular context and moment to ask your question, you either betray too much of your own perspective in the formulation you use or, in the heat of the moment, are lost for words. This is why it is useful to pose topic headings in the form of questions so that you do not have to do so much thinking on your feet. In general you should ask open questions, not only in the sense of avoiding questions that can be answered with a simple yes or no (unless you follow this up with a 'Can you say a bit more about that?'), but also avoiding formulations that could be interpreted as prescriptions for, or prohibitions on, what can be talked about - unless you consider that the situation or topic merits you positioning yourself more clearly.
Second, while these recommendations may seem daunting, all becomes much clearer and easier when you do a practice interview, perhaps with a friend with whom you feel at ease and who can give you frank feedback on the content and process of the interview. This helps to identify and iron out problems with the interview schedule and with the recording equipment (so that you remember to switch it on, you know where to position the microphone, you know not to have the machine on autoreverse so that it records over the first side again, etc.). Not least, you will gain a lot of confidence from the experience, even if you are also made acutely aware of the demands made on you as interviewer. These include the capacities to focus in parallel on listening intently to what your interviewee says, reflecting on how this relates to your interests, preconceptions and schedule, and working out what to say and when to say it.
The sample analysis below reflects the focus so far on reflexive issues. It draws on aspects of all models identified above, but is most influenced by feminist and postmodernist approaches. In looking at this thematic analysis it is important to be clear that, even when carefully selected, abstracted from its original context and juxtaposed with other examples, as are the extracts below, they do not 'speak for themselves'. Meaning inheres not only in the text but in our construction and reading of it: despite the process of selection and interpretation in the preparation of this illustrative material, the analysis is inevitably selective. We are going to illustrate features of interviewing practice through a focus on instances of metacommentary (commentary on, in this case, the research process) occurring within some extracts from interview transcripts. At the outset we need to identify the questions in relation to which the analysis is structured, present the rationale for the material used and introduce the material itself
The questions selected here concern the visibility of rules governing the structure of the interview, which here are exhibited through moments of role-shifts and changes between interviewer and participant. Through this analysis we aim to illustrate how participants hold and use their positions within the research relationship. That is, they are neither passive nor unknowing about their positioning, but rather use this to achieve specific outcomes within the interview. We would not claim that these examples are routinely found within interview transcripts, but they are certainly not unique. The fact that these examples originate from interviews in which all the interviewers knew their informants beforehand (and indeed presumably selected them as willing to talk, and to talk about the selected topic) perhaps enabled the expression of what is often either non-verbally or indirectly communicated, or commented on off-tape. In line with current reflexive work on research practice, the focus here is on making visible the researcher within the research context. Correspondingly, for purposes of exposition we are going to exaggerate the reflexive ethos to focus not on the individual interview, or interviewer, but on what these extracts can tell us about research processes (which also includes what we are doing with them here).
Construction and selection of interview material
The extracts below are drawn from three second-year undergraduate practicals where students were conducting a single semi-structured interview supervised by one of us on a topic of their choice. As is appropriate research practice, they negotiated a research contract with the participant before the interview itself, which included ensuring that the transcripts would remain anonymous and would be read only by them and the markers of the work (of which, as the supervisor, I was one). Students were encouraged to discuss the interview process with participants to promote reciprocity and feedback. Similarly, these are the conditions under which I am drawing on these interviews. I am reproducing these extracts with the students' permission and with that of their informants. The fact that I did not conduct these interviews, and moreover am presenting a necessarily selective and motivated analysis, is an issue that I flag here to take up later.
It is a moot point whether the material for analysis here is the text of the interview or the interview itself. Frequently researchers take as their record, as their 'raw material' for analysis, the transcript of their interview. This is belied by the fact that a transcript is: (a) an impoverished record, a key stopping point on the road of progressive removal from encounter, to aural representation (on tape), to written representation; and (b) therefore a selective/constructive representation, as highlighted by the variety of transcription notations that embody their own assumptions, whether through spatial arrangement (Ochs 1979) or through levels of detail about what is important (discussed briefly in Stubbs 1983, and extensively in Tedlock 1984).
Furthermore, not only is the production of a transcript also part of the research process, but the interviewer brings to the transcript her experience and memory of the interview. Here too it is appropriate to articulate impressions and perceptions of emergent issues and feelings, preferably as soon after the interview as possible, and in any case before starting on the analysis. These 'field notes' can become a resource, both in informing the analysis and in reminding you of what assumptions you brought to that analysis - which may or may not be borne out. Given all this construction and selection in the process of organisation and synthesis that is necessary for an analysis, we should also question what is excluded, suppressed. The material presented here for analysis is of a variety that in many accounts would be sanitised away, deemed bad research, embarrassing intrusions of the personal or even lapses of interviewer control. These are precisely what make them interesting and important to analyse as indicative of the implicit rules of research, rules which become more apparent in their infraction.
A 'thematic analysis' is a coherent way of organising or reading some interview material in relation to specific research questions. These readings are organised under thematic headings in ways that attempt to do justice both to the elements of the research question and to the preoccupations of the interviewees. I will start by presenting each extract separately and then move on to elaborate connections and contrasts between them in relation to the themes of power relations in research, the power of the participant and shifts in positions. The texts are longer here, for illustrative purposes, than would normally be presented within the main text of a report. The transcription and terminology (I, interviewer; R, respondent; P, participant) are as in the originals. Clearly these terms reflect different ways of positioning the person being interviewed, and precisely because of this I have chosen to retain terminology from the original transcripts. Line numbers refer to those parts of the transcripts presented here.
Presentation of text and general comments
An interview by a young male student with a woman friend on the topic of formative influences on occupational choice.
I I: relax - this is nothing that will be judged
2 R: I'm not looking at the questions
3 I: All I want is to find out a few things concerning your family,
4 educational, friends - aspirations, motivations etcetera
5 R: okay
6 I: and I'll give you feedback on it - later on
7 R: not now
8 I: no - you can have anything you have said which you do not want
9 to be disclosed erased
10 R: aha
11 1: you can also have anonymity - you can choose your pseudonym
12 if you want to
13 R: oh fine
14 I: as long as it isn't (interviewer's name)
15 R: [laughs]
16 I: oh yes I'll be taking a few notes which have nothing to do with
17 your answers during the interview - is this okay
18 R: you'll be taking notes - I see - okay
19 I no
20 R: so you're going to be watching me as well
21 I: I'll record anything that's interesting - relevant
22 R: I'll put my hands behind my back then...
23 R: ... but that's life - as long as I learn from my mistakes - what
24 was the question again - sorry
25 I: it's okay you've answered it
26 R: good - stop looking at my legs [laughs]
27 I: erase
28 R: sorry I just wanted to say that - sorry
29 I: Hmm - let's continue - okay - and cover your legs...
It seems that the prior friendship enables the respondent (R) to comment on the prevailing rules structuring the interview by highlighting the assumptions of control implicit within the interviewer's (I) framing of the interview: the feedback is 'not now' (1.7) but later. While it is important not to minimise the ways interviews can become opportunities for sexual harassment (see later), it is also possible to read this exchange as R asserting in a mock threatening way her power to evaluate and censor the interview material. She comments on being scrutinised both verbally ('you'll be taking notes - I see - okay', 1.18) and visually ('so you're going to be watching me as well', 1.20). Throughout this extract there is an air of what I read as amicably sarcastic compliance with clear suggestions of how provisional this may be ('aha', 1.10) and how the resistance can enter into what is available to be recorded even before the explicit rights of later erasure can be exercised ('I'll put my hands behind my back then', 1.22). R's apparently wilful misinterpretation of I's attempts to assert his position as interviewer and hers as respondent occurs even before the critical moment of embarrassing I with 'stop looking at my legs' (1.26). Significantly, this is immediately after R's indication that she had become so absorbed in her narration that she had forgotten the interviewing context (1.23-4), and so this can be read as a correlative disengagement. It is possible to read the two times (1.8 and 1.19) I says 'no' more as dismay and suppressed repudiation of R's resistance rather than a refusal of any specific feature of it.
A male interviewer of a male acquaintance on his political involvement with animal rights groups. In contrast to the tense atmosphere of Extract 1, I read this extract as a more reflective and collaborative exchange in which the interview gravitates to become an interview about interviewing.
1 I: Do you think it was a wise move for me to get into psychology?
2 R: I thought you were supposed to be interviewing me
3 I: what's the difference
4 R: I don't know
5 I: do you think the person being interviewed can get as much or
6 more out of the interview as the interviewer
7 R: yes of course, they wouldn't agree to be interviewed otherwise.
8 People enjoy being asked questions about what they are interested
10 I: But the police don't ask if you agree to be interviewed
11 R: No, but you don't have to say anything
A woman interviewer of a woman friend on the topic of the importance of friendship.
1 P: .. I know this is difficult for you (I's name) but you did ask me
3 do this and I couldn't possibly talk about friendship without
4 talking about you
5 I: right OK then (laughs)
6 P: Right, it's very very important to me but erm it's quite difficult
7 actually because I'm talking to you actually am I talking to you,
8 the person or you an interviewer?
9 I: I think we can say you can talk about me as a person it can't be
10 objective it's subjective anyway
11 P: It's really weird to be actually talking about you erm it s not
12 embarrassing is it? I have to do it.
13 I: a bit but you will have to be clear for people reading it where we
14 met and that
15 P: I can talk away from you if you want, right I will sort of Yea
16 the person I'm talking about her name is...
17 I: do you think it [the interview] could be improved?
18 P: no I don't because I wouldn't be, if I had somebody, if the
19 interviewer was somebody who, a person I didn't know, I really
20 don't think I would convey my feelings with very many people
21 erm in actual fact I would only do this interview for you I
22 wouldn't do it for anybody else I certainly wouldn't talk about
23 friendship the topic or whatever with anybody else it would
24 have to be a close friend
25 I: how do you feel about all of this being seen by others?
26 P: How do I feel about that, I don't mind really I don't mind
27 I: There is so much material on there and I just feel I'm going to use
28 it for adverse means in a sense and I wouldn't like it er
29 P: No no I don't mind the reason I don't mind is that whatever is
30 recorded on there whatever is transcribed on to paper will be in your
31 hands that's what's important to me No it doesn't matter actually
32 whatever you come up with it really doesn't matter. But
33 nevertheless it's in your hands you're taking care of it it's your
34 baby I'm entrusting my feelings my thoughts and feelings to you
35 which I already know I feel very comfortable with anyway this makes
36 no difference in a sense it really doesn't.
This interview moves from an emphasis on the meaning of friendship in general to a focus on support, and supportive friends in particular. The participant (P) uses the interview to tell the interviewer (I) about the importance of I's friendship to her; that is, she uses the new set of positions within their relationship combined with the public nature of the encounter to comment on the value of her relationship with I. The positions are explicitly marked where P asks 'you the person or you an interviewer' (3.7-8), which sets up a corresponding set of positions for her as both person and respondent. She constructs I as a research object to include her as proper material for the study by 'talking away from you' (3.15) and describing her in the third person ('the person I'm talking about her name is...', 3.16).
Having introduced the material, we can now use four themes by which the extracts as a whole can be viewed in relation to the analytic questions identified earlier. You might want to consider what your choice of themes might be before reading further.
Theme 1 Using the position of respondent
In Extract 1, R can be seen as asserting her power as respondent in a context in which I is, in this case quite explicitly, dependent on R for the completion of his work. In addition, the embarrassment ('erase', 1.27) over looking at, or being positioned as looking at, R's legs highlights the covert sexuality, in this case heterosexuality, of the research encounter. In this research encounter, the official power of definition and interpretation coincides with that of the active viewer and evaluator. These culturally masculine positions were here held by a male researcher. R demonstrates her power over the interpretation of the encounter by invoking the framework of harassment, thus enlisting a range of judicial and disciplinary practices. One reading of this exchange, then, is that by making the potential sexual exploitation of the encounter explicit, the female research participant both comments on this convergence of gender, sexual and interview power relations and resists being positioned as a passive subject of the research gaze by using the structures of the research (question-asking, recording) to turn the researcher himself into an object of scrutiny.
In Extract 2 it is the interviewer who sanctions the commentary about the interview process with 'do you think it was a wise move for me to get into psychology?' (2.1). The explicit discussion of interviewing roles follows from I's assertion of, and departure from, the role of interviewer by asking a question, but of a reflexive form that transgresses, transcends or reverses the current interviewing relationship.
In Extract 3 the additional positions made available by the interview appear to be used by P to enable her to thank I and express her appreciation of I as a friend. Using the rules of the interview, not only is P required to hear this (owing to the injunction upon I to listen to and respect what P says), but it is also justified within the rules of interviewing, owing to the norm of frank disclosure conventionally assumed of research participants.
Theme 2 Respondent-initiated reflexive commentaries
In these extracts, although in different ways, the respondents initiate the reflection on the research process. In Extract 1, R exploits the novice I's highhandedness in reminding her of the research contract, which he appears to inform her of, rather than negotiate. She therefore frustrates I's attempts to assume the position of authority as the one who could but will not judge (1.1), who defines what is relevant to record (1.21), who has defined the questions (1.3) to be answered, and who has the authority to provide feedback (1.6). In Extract 2, R challenges I's assumption that he can change the roles or rules of the interview with his 'I thought you were supposed to be interviewing me' (2.2). In Extract 3, in an interview about friendship, it is significant that P raises the reflexive issues about her friendship with I. This emphasizes how commentaries on the research process are by no means the prerogative of interviewers. Rather such cues and feedback are always present even if unarticulated, and form the infrastructure of research.
Theme 3 Public nature of the account
It is clear that the fact that these are interviews, rather than conversations, constitutes key structural conditions for the accounts generated (and the division between the 'public' and 'private' itself has a history). R's resistance in Extract 1 to the interviewer-respondent relation is exercised precisely through the public nature of their encounter. I's reaction to her request to 'stop looking at my legs' is in terms of his awareness of the tape-recording (he says 'erase', 1.27), while R reasserts her desire to record this (literally reinscribes it) with 'sorry I just wanted to say that - sorry' (1.28). R therefore enlists the broader structure of scrutiny hinted at earlier by I's prohibition that R should use his name as a pseudonym (1.14), which would place him as doubly subjected to the interview (as interviewer and respondent). (In psychoanalytic terms we might see this as a return of the repressed with a vengeance!)
Within Extract 2, I initiates the discussion over who benefits from the research ('do you think the person being interviewed can get as much or more out of the interview as the interviewer', 2.5-6). I also develops R's response in terms of people enjoying talking about what they are interested in (2.8-9) by importing a legal-political interpretation of people's 'interests' (in police interrogations) (2.10). The comparison between legal/coercive regulation and the structure of the interview is both taken up and refused by R in his reminder of participants' voluntary involvement as conditional on personal engagement or relevance, as exhibited by the strategies available for non-compliance within even the most formal structures of interrogation (2.7-11). It may be appropriate that I defers to R over the success of this strategy; the probability that he has had to exercise this is rendered more likely by the topic of the interview. In Extract 3, the public nature of the account as well as the topic (friendship) is a prerequisite for P expressing and reasserting her trust in I. Both parties refer to a (temporally distanced but very present) audience ('you will have to be clear for people reading it', 3.13, and 'how do you feel about all of this being seen by others', 3.25).
Theme 4 Past/future relations
In all three extracts, the shifts in I-P positions are achieved by reference to, and in relation to, their relationships outside the interviews. In 1, the teasing tenor of the interview suggests that R wields her power through the prior relationship she has with him. In 2, I marks permission to digress from the previously agreed topic of the interview by asking R a question that invites judgement of I on the basis of their prior relationship. He asks R for his perception of the rules of the interview with 'what's the difference' (2.3). The past relationship is alluded to by reference to the choice to be interviewed. As interviewer 2 commented to me after reading this analysis, either his mistrust of psychology is well-placed, or else, as a prior acquaintance, the interviewer is a particularly clever policeman to trick him into thinking he has a choice. In Extract 3, by the end of the interview the research encounter is incorporated into the structure of the existing friendship as not only a test of the relationship ('it's in your hands... I'm entrusting my feelings my thoughts and feelings to you', 3.33-4), but as an expression of it ('I would only do this interview for you', 3.22, 'this makes no difference in a sense it really doesn't', 3.35-6), and by the end the interview itself is in the past, as 'there' on the tape (not here and now), in 'there 15 50 much material on there' (3.27), and as a record that can be used or abused.
In this section we locate the analysis within an account of its production, including relevant constraints, limits and possibilities wrought by our position as analysts. These are issues that might be regarded as the context of the research, and this 'context' (in the sense of what accompanies and constructs the text) can be divided into ten points.
Records - what's lost (and gained)
The tendency to confound interview with transcript discussed earlier is in some senses clarified by the fact that the analysis presented here is based on transcripts of interviews which I did not conduct or transcribe. Paradoxically, the fact that I did not conduct the interviews puts my position closer to that of other readers in both generating and justifying the analysis. What this brings home is the incompleteness and partiality of interpretation (see below), qualities intrinsic to this kind of research but made more manifest in these circumstances. I have reproduced the extracts using the notation adopted by the original interviewers/transcribers. The use of conventional punctuation and absence of explicit transcription codes in part reflected their concern about presenting to informants a formalised and inaccessible representation of the interview that might compound further the potentially alienating effect of seeing spoken language written down in all its (in terms of written codes) untidiness (Stubbs 1983). But these extracts would perhaps have gained from a more systematic coding (see the Appendix for a simplified notation suitable for most thematically based analyses).
A common reaction to analyses of the kind offered above is that the material has been misinterpreted or overinterpreted, manipulated to produce meanings that were not 'originally' there. It is certainly true that the process of analysis, including the shifting representational forms that the interview/text assumes, does provide new vantage points from which to interpret, and that, as Stubbs (1983) points out, close scrutiny of a transcript can magnify tensions or aggressive elements within the text. However, acknowledging this does not discredit the analysis offered; rather it supports the suggestion that this is one of multiple ways of reading the texts. Clearly if the reading is mine alone, that is, if this reading is not recognised by others, then its credibility is undermined. Similarly, the reactions of the original interviewers and participants in the extracts are important. Depending on the model of research adopted, disagreement would invalidate the reading (within an ethnographic model) or be interpreted within it (since from a postmodernist stance it could be argued that there may be particular investments in refusing the interpretation; Opie 1992). Here it is worth saying that the original interviewers' reactions to this analysis were very favourable, and no reservations were expressed about the readings outlined above, nor any alternatives put forward. In terms of additional points, the interviewer in Extract 1 wished to have further emphasized how much is lost in the transition from tape to transcript. Feedback from interviewer and interviewee 2 confirmed the view of the interview process as an elaboration of mutual agendas, and indicated a view that this should be recognised and explicitly structured into interview procedures.
All this emphasizes the constructed and inexhaustive nature of the analysis. These features sometimes make this kind of research both frustrating and dissatisfying, since a common response is to feel acutely the partiality (in the double sense of incomplete and motivated) of interpretation. Nevertheless, this too is an instructive reflection on the research process. It is helpful here to explore the infinite regress and incoherence of the fantasy of a complete and authentic original record (Why use audiotape, why not use videotape? But what about that space behind the camera?), which parallels that of the complete interpretation. Rather, we should accept the uncertainty of unfinished analysis as an index of the arbitrary limit imposed by writing up. In principle the research process could continue almost indefinitely, both in the sense of exchanging readings and reactions between interpreter and participants, and in the analyst's shifting perceptions of their interpretations (and clearly some such mutual discussion and reflection with participants is good research practice).
Intentionality and multiple readings
There are clearly other ways of interpreting the extracts above. In fact it is often useful to present and develop alternative readings and explore the different conclusions they indicate. For my purposes I have selected one reading, but in drawing attention to its provisionality two issues should be noted. First, acknowledging the multiplicity of readings (or accounts of the account) does not mean that all readings are equal. Otherwise we are stuck in the quagmire of relativistic nihilism which disempowers us from using the research to say anything (see Bhavnani 1990; Burman 1990). There may be good reasons for privileging the reading or account of the research participants, particularly if their 'voice' is that of a disadvantaged or under-represented group (but see problems with this in Chapter 8). Second, and this is particularly relevant for the kind of meta-analysis I have presented here, it is important not to equate the reading with either the intentions of participants or their intentions in the interview. The analytic questions driving the analysis offered here were concerned with research process and not individual opinions, so the focus was correspondingly on implicit features of the rules or frameworks structuring research rather than on what each party said or did. In this sense other analyses of this material could be made in relation to different questions from those explored here.
Selection of material
A reflexive account could include examination of my own motives in presenting this material. Perhaps I am aiming to demonstrate something about my teaching practice through the interviewing competence of my students. Perhaps I am avoiding subjecting my own interviewing and analytic practice to critical scrutiny (but see Burman 1992a, b). A more standard criticism here would question the validity of generating this analysis from interviews conducted on different topics and by different interviewers. However, once again this approach can be defended by recalling that the analytic issues do not require absolute comparability of text (indeed such a notion is questionable, see Chapter 1), and in fact may benefit from the variety of positions adopted.
In more typical circumstances, the researcher has privileged knowledge both of the participants and of the experience of conducting the interviews. In this case the extra knowledge I bring to bear on this material is my acquaintance with the interviewers, the history of having supervised (and marked) their interviewing research, and therefore my access to the larger transcripts from which the above extracts were selected. Again, these differences in position from which the extracts can be read have to be acknowledged rather than erased, and can perhaps be put forward either to explain differences in or to fix interpretation (Burman 1993).
A legitimate question that should always be posed (both in conducting and in evaluating research) is whether the participants have been exploited, that is, whether their psychological or material conditions worsened through their involvement in the research. In this case, permission was gained from all parties to use the material, and they were consulted over the interpretations drawn. As discussed in the last three chapters of this book, issues of exploitation go beyond notions of 'informed consent' to include the use made of the research.
Effects of prior relationships
These arise in the form of the interviewers' prior knowledge of their participants from other than research contexts, and of my prior knowledge of the interviewers from a teaching context. In the first case it seems likely that this facilitated greater disclosure and reflexive commentary, as well as constituting the preconditions for some of the themes identified in the analysis. The impact of the second issue is difficult to evaluate but clearly does figure; quite how helpful or otherwise this is will depend on how persuasive you as readers find this account. We are not, however, suggesting that people should only interview their friends, but rather we want to high-light how the prior relationship (of acquaintance, or non-acquaintance, which is also a relationship) enters into the structure and content of the encounter.
Danger of fetishing particular strategies
Similarly, the analysis here should not be read as recommending particular interviewing devices in order to promote either equality or reflection on research processes. Used as techniques these can work to assert interviewer authority in indirect ways (see Walkerdine 1988: 53-63; Burman 1992a, for examples from research with children). The examples presented here arose spontaneously, and as initiated or at least taken up by the 'respondents'. The analysis here does not (and cannot) illustrate how interviews should be done, but rather offers suggestions about what to look for, and how to think about what happens in interviews. There are no techniques or analytic procedures that escape the dangers of exploitation. Hence it is important to structure consultation and feedback over the interpretation of transcripts with participants.
Interpretive stance/countertransference issues
Given the nine points above, it is clear that the analyst brings to his or her analysis a range of different identifications and responses. In these extracts I identify as interviewer struggling to democratise the research process and anxious about how to make sense of the material (perhaps anyone who has engaged with semi-structured interviews can identify with the sense of responsibility that the interviewer comments on in Extract 3, line 25). I also hold multiple positionings and identifications arising from structures of gender, class and so on, which inevitably enter into the particular analysis formulated. Where personal reactions or investments play an important part, these can be treated as a resource for, rather than 'interference' within, the analysis (for an example of this see Marks, in press).
Discussion and assessment
The analysis presented here is based on unusual material. These examples may not be everyday but they are also not unique, and are consistent with the aim of highlighting the visibility of the researcher rather than the researched. Qualitative analysis elucidates phenomena that would be missed or dismissed by other methods. Just as the exception can prove the rule, so exceptional or incidental instances can function to highlight structural dynamics that underlie research encounters.
Part of what is so arresting about this material, in particular Extract 1, is that it refers quite literally to what is absent or lost from records of research: all too often we analyse interviews as disembodied voices but interviews are interactions between embodied people. In the example discussed here, R made this visible by protesting against I's gaze. Issues of sexuality within research encounters are rarely commented upon in accounts of research, but this example illustrates aspects that must always be present, even if suppressed. Where the gender and researcher-researched positions are reversed, that is where men are interviewed by women, issues of participant resistance have been documented as being acted out in the form of sexual harassment of researchers (see also McKee and O'Brian 1983; Reynolds 1993). It seems that where traditional power relations are departed from, where men are positioned as subordinate within the researcher-researched relation, for example, this structural power relationship becomes evident through their attempts to subvert it. The disjunction or exception highlights the presence of the rule.
We should draw attention to the concept of power that has informed this analysis. Although it is concerned with instances of 'respondent' resistance or assertion, this analysis does not imply that 'respondents' have greater power than 'interviewers'. Rather, following a Foucauldian perspective, power is not conceived of as a unidimensional quality that is possessed or lacked. This analysis has illustrated how particular possibilities for assertion and resistance are produced through the structures of the research encounter. Specifically, it has been argued that: (a) once consent has been given, 'respondents' or 'participants' are not passive parties to the question-answer interviewing structure; (b) they can assume a range of strategies to resist that positioning; (c) they intervene in and comment on the research process, as well as being the research focus; and (d) they achieve both joint and separate goals through their participation in the research.
In evaluating the conduct and analysis of interviews, there are continuities between thematic and textual approaches, and similarities (although cast within rather different terminology and corresponding philosophical basis) between the account here and 'grounded theory' as elaborated by the criteria identified by Henwood and Pidgeon (1993: 24-7) of 'keeping close to the data; theory integrated at diverse levels of abstraction; reflexivity; documentation; theoretical sampling and negative case analysis, sensitivity to negotiated realities, transferability.' While most of the limits of this approach have been addressed in the reflexive analysis, it should be noted that interviewing is time-consuming, absorbing and suited to a study with a restricted number of interviews in order to keep the transcription and analysis of material manageable, and to do justice to the material generated. In terms of problems, the multiple frameworks informing semi-structured interviewing can make the approaches appear atheoretical or intuitivist. A disciplined analysis of the presuppositions guiding all stages of the research does much to ward off this criticism. The final analysis, however, lies beyond the account here in the interpretations made of this by you.
Suggested transcription notation:
(2) two second pause (number indicates duration)
(xxx) indistinct/doubtful transcription
Word underline emphasis
Potter and Wetherell (1987: 188-9) suggest a slightly more complex notation, which in turn is a simplified version of Sacks et at (1974).
Gubrium, J. and Silverman, D. (eds) (1989). The Politics of Field Research. London: Sage.
Henwood, K. and Pidgeon, N. (1993). 'Qualitative research and psychological theorising', in M. Hammersley (ed.) Social Researching: Philosophy, Politics and Practice. London: Sage.
Nencel, L. and Pels, P. (eds) (1991). Constructing Knowledge: Authority and Critique in Social Science. London: Sage.
Steier, F. (ed.) (1991). Research and Reflexivity. London: Sage.
Bhavnani, K.K. (1990). '"What's power got to do with it?" Empowerment and social research', in I. Parker and J. Shotter (eds) Deconstructing Social Psychology. London: Routledge.
Bingham, W. and Moore, B. (1959). How to Interview. New York: Harper International.
Burman, E. (1990). 'Differing with deconstruction: a feminist critique', in I. Parker and J. Shotter (eds) Deconstructing Social Psychology. London: Routledge.
Burman, E. (1992a). 'Feminism and discourse in developmental psychology: identification, subjectivity and interpretation'. Feminism and Psychology, 2 (1), 45-60.
Burman, F. (1992b). 'Identification and power in feminist therapy: a reflexive history of a discourse analysis'. Women's Studies International Forum, 15 (4), 487-98.
Burman, F. (1993). 'Beyond discursive relativism: power and subjectivity in developmental psychology', in H. Stam, L. Mos, W. Thorngate and B. Caplan (eds) Recent Trends in Theoretical Psychology, Volume III. New York: Springer Verlag.
Figueroa, H. and Lopez, M.M. (1991). 'Commentary on Second Discourse Analysis Workshop/Conference'. Paper presented at the Second Discourse Analysis Workshop/Conference, Manchester, July.
Gubrium, J. and Silverman, D. (eds) (1989). The Politics of Field Research. London: Sage.
Henwood, K. and Pidgeon, N. (1993). 'Qualitative research and psychological theorising', in M. Hammersley (ed.) Social Researching: Philosophy, Politics and Practice. London: Sage.
McKee, L. and O'Brian, M. (1983). 'Interviewing men: "taking gender seriously" ', in F. Garmarnikow, D. Morgan, J. Purvis and D. Taylorson (eds) The Public and the Private. London: Heinemann.
Marks, D. (forthcoming/in press) 'Gendered care dynamics and the structuring of group relations: child-professional-parent' in F. Burman, P. Alldred, C. Bewley, C. Heenan, D. Marks, J. Marshall, K. Taylor, R. Ullah and S. Warner, Challenging Women: Psychology's Exclusions, Feminist Possibilities. Buckingham:
Open University Press.
Mishler, F. (1986). Research Interviewing: Context and Narrative. Cambridge, MA:
Harvard University Press.
Nencel, L. and Pels, P. (eds) (1991). Constructing Knowledge: Authority and Critique in Social Science. London: Sage.
Ochs, E. (1979). 'Transcription as theory', in F. Ochs and B. Schieffelin (eds) Developmental Pragmatics. New York: Academic Press.
Opie, A. (1992). 'Qualitative research, appropriation of the "other" and empowerment', Feminist Review, 40, 52-69.
Piaget, J. (1929). The Child's Conception of the World. London: Routledge and Kegan Paul.
Potter, J. and Wetherell, M. (1987). Discourse and Social Psychology. London: Sage. Reason, P. and Rowan, J. (eds) (1981). Human Inquiry: a Sourcebook of New Paradigm Research. Chichester: Wiley.
Reynolds, G. (1993). '"And Gill came tumbling after": gender, emotions and a research dilemma', in M. Kennedy, C. Lubelska and V. Walsh (eds) Making Connections: Women's Studies, Women's Movements, Women's Lives. London:
Taylor and Francis.
Sacks, H., Schegloff, F. and Jefferson, G. (1974). 'The simplest semantics for the organisation of turn-taking in conversation', Language, 50, 697-737.
Spender, D. (1981). 'The gatekeepers: a feminist critique of academic publishing', in H. Roberts (ed.) Doing Feminist Research. London: Routledge and Kegan Paul.
Spradley, J. (1979). The Ethnographic Interview. New York: Holt, Rinehart and Winston.
Steier, F. (ed.) (1991). Research and Reflexivity. London: Sage.
Stubbs, M. (1983). Discourse Analysis: the Sociolinguistic Analysis of Natural language. Oxford: Blackwell.
Tedlock, D. (1984). The Spoken Word and the Work of Interpretation. Philadelphia: University of Pennsylvania Press.
Walkerdine, V. (1988). The Mastery of Reason. London: Routledge.
I am very grateful to the interviewers and participants whose talk and work I draw upon in this chapter, in particular for their prompt and supportive feedback on earlier drafts.