Chapter 5
niervsewin
Although some social research relies purely upon observational techniques and some can be done without ever stepping outside a library, many types of research rely, to a greater or lesser degree, upon asking people for factual information, or questions about what they do and do not do, or about their beliefs, attitudes, aspirations, experiences and feelings. Interviewing people, whether for survey, case study or ethnographic research, presents the researcher with a number of practical problems, in particular, interviewers often face difficulties in:
" Obtaining accurate and truthful responses to relatively closed questions, such as 'How often?', or 'How many?'
° Obtaining fall and sincere responses to open ended questions such as 'How did you feel about x?', 'Can you tell me what happened when y?'
e Focusing the interview, that is, getting people to talk about the issues which concern the researcher
Different methodological traditions emphasise different problems and different ways of dealing with them, and this chapter begins by looking at advice to interviewers from orthodox, qualitative and feminist methodologists. It then considers the techniques employed by Kinsey, Freud and Scully to highlight some of the issues raised by interviewing in the real world. This allows us to consider a methodological problem which is rarely dealt with in textbooks on interviewing. For while all manner of methodologists have a lot to say about how best to get people to give truthful, or full and sincere replies, less is said about how a researcher can tell whether an interviewee is telling the truth or being completely frank and open. Yet this issue is vital to the process of interviewing and to the reliability of the data that is collected.
Interviewing    117
THE ORTHODOX APPROACH TO ASKING QUESTIONS
Central to positivist philosophy is the assumption that there is a world of hard facts which have an existence independent of their social context and separate from the meanings that people attach to them. Take an example from the natural world. Mount Snowdon may strike someone from the Netherlands as being very tall, whilst to someone from the Himalayas, it may appear to be quite small. But though different people's subjective perceptions of the size of this mountain may differ, in reality, Mount Snowdon is a definite and particular height. The natural scientist's aim is to develop neutral instruments to measure this objective reality, rather than relying upon the subjective perceptions of individuals. Positivists hold that the same approach is necessary in the social sciences. We do not want to rely on the individual interviewer's subjective perceptions of how sexist or racist an interviewee is, for example, since the interviewer's subjective judgements are unreliable. The social sciences need a method which can strip away the subjective meanings that people attach to behaviours, ideas and events, and discover objective truths about the social world. It is therefore necessary to apply, as closely as possible, the methods of the natural sciences to the study of the social world. This is no easy task, especially when the research relies upon asking people questions.
When natural scientists investigate the molecular structure of a piece of metal, there is no danger that the subjective perceptions and beliefs of the piece of metal will interfere with and pollute the research process. Natural scientists need not concern themselves with the metal's reaction to the lab technician. The metal is hardly likely to refuse to be examined by someone because it does not like the look of them. Neither do natural scientists have to worry that the metal will attempt to deceive the lab technician in order to appear better endowed with molecules than it really is. The social scientist investigating the attitudes and behaviour of people, however, is dealing with conscious, purposive actors. In setting up interviews, the researcher is setting up a series of social interactions between interviewers and the human subjects of the research. Both interviewer and respondent bring to the interview a set of subjective beliefs, expectations, values and so on, which could potentially obscure or distort the truth. To give a rather obvious example, one of the things Kinsey asked male respondents was to estimate the length of their erect penis. Over 17 per cent of white college males estimated the length to be more than 7 inches, with four men even claiming that their members extended more than 10 inches (figures taken from Gebhard and Johnson
118    Methods, sex and madness
1979: 116). Had this question been put to them by a woman they found attractive, rather than Kinsey or one of his male associates, it is possible that their replies would have stretched the bounds of credulity still further. In other words, the interview is a social encounter, and how the respondent answers questions will depend to some degree upon what the respondent and interviewer think and feel about each other. For positivists, this raises the central methodological problem so far as interviewing is concerned. How can the researcher be certain that the respondent will give accurate and truthful information, rather than trying to please or impress the interviewer? How can researchers ensure that the data culled through interviewing is not coloured by the interviewer's subjective perceptions of the respondent and the respondent's subjective perceptions of the interviewer? Advice on interviewing techniques in orthodox textbooks reflects this central concern.
Whether interviews are being used in survey or case study research, orthodox textbooks hold that the researcher must take certain steps to avoid the bias that could arise from interviewer-respondent interaction. To begin with, the researcher must ensure that despite the fact that interviewers and respondents each have their own individual personality, history and mannerisms, each interview is standard and identical. An interviewer asking questions about sexual behaviour should ask every respondent the same questions about sexual contacts with animals, not omit these questions out of a sense of delicacy when interviewing nuns, for example. If nuns are not given an opportunity to answer these same questions, then the information from them will not be comparable with the information from the other respondents. The data must be gathered in a standard way. Similarly, in large-scale research where a number of different interviewers are used, the researcher must be wary of what is known as 'between interviewer variance'. Imagine two different interviewers collecting data for a survey on sexual behaviour. If the first asks someone how many times a month they have sex, and ticks the 'don't know' box as soon as the respondent replies 'I don't know', while the second gives all sorts of prompts to the respondent ('Is it about once a month, or maybe twice? With most people it's about two or three times, would you say you did it more or less than average?') and in this way finally extracts a figure, then the data culled by the two interviewers is not comparable. It was not gathered in a standard way.
Orthodox methodologSsts emphasise this need for standardisation. Regardless of whether a tightly structured, formal interview schedule or a very loosely structured, topic-based interview is employed, the
Interviewing    119
important thing is to provide each interviewer with very clear and explicit instructions as to the questions to be asked or issues to be covered, the prompts that can be used, and the way in which responses are to be coded. In this way, the researcher ensures that even if every respondent were to be interviewed by a different person, they would each have a standard interview, coded in a standard way. To achieve this end, it is essential to provide interviewers with detailed training in how to approach respondents, how to tell respondents about the research, how to ask questions in the same way, how to code and record responses. Through standardisation, the researcher ensures that each respondent experiences an identical interview and that the data gathered through interviews is standard and comparable.
For large-scale survey research, interviewer selection is seen as vital. Orthodox textbooks often stress the need to recruit interviewers whose personal characteristics will not interfere with the subject's responses and to train them to undertake interviews in a neutral, professional fashion. The issue of interviewer selection ties in with what is, for many orthodox methodologists, the most insoluble paradox associated with interviewing. On the one hand, there must be a rapport between interviewer and interviewee. If respondents find the interviewer hostile, or unpleasant in some way, they are unlikely to co-operate with the interview, far less reveal any intimate truths. But on the other hand, researchers have to guard against bias that may arise as a consequence of subjects getting on with the interviewer too well. If there is too much rapport, the respondent may try to please the interviewer by saying what he or she thinks the interviewer wants to hear, rather than telling the whole truth. Moser and Kalton (1971) state that:
There is something to be said for the interviewer who, while friendly and interested does not get too emotionally involved with the respondent and his problems. Interviewing on most surveys is a fairly straightforward job, not one calling for exceptional industry, charm or tact. Pleasantness and a business-like nature is the ideal combination.
(1971: 286)
Questions must be asked as well as phrased in a neutral way, because if interviewers make it plain through facial expressions, intonation or verbal comment that certain views are either desired by, or unacceptable to them, the respondent is likely to modify his or her responses. No one would wish to confide details of a homosexual experience to an interviewer who was plainly homophobic, for example. The behaviour of the interviewer can thus lead to bias. Moser and Kalton hold that the way to
120    Methods, sex and madness
reduce such bias is to carefully train the interviewer 'generally to deport herself in a way that is least likely to influence the respondent's answers'(1971: 272).
The concern with standardisation also leads orthodox methodologists to stress the need to control the subjectivity of the respondent. Because the respondent is a conscious, purposive actor, he or she can distort the interview by asking questions or making comments. If the interviewer responds to the respondent, striking up a conversation about the research and related issues, then there is no hope of ensuring standardisation, since that particular respondent will be provided with information that is not available to others, and furthermore, the interviewer's replies to questions might bias the respondent's future responses. For these reasons, orthodox textbooks hold that it is essential for the interviewer to retain complete control during the interview:
[T]ight control is a central goal of the interviewer. The interviewer must take complete charge of the interaction, including such things as where people sit, when the interview begins and ends, what topics are covered, when they are covered, and so on. Losing control of the interview is almost always a methodological disaster that terminates useful data gathering.
(Hessler 1992: 137)
The focus and scope of the interview must be controlled by the interviewer, and this both requires and implies a firmly hierarchical relationship between interviewer and respondent. The interviewer must be pleasant in order to secure co-operation, but must leave the respondent in no doubt as to who is in charge of the situation. The goal of standardisation further requires the interviewer to remain as neutral and as detached from the respondent as possible, whilst simultaneously maintaining a good rapport. Interviewers should politely but firmly refuse to engage in conversation with the interviewee, by saying something like 'I am much more interested in hearing what you have to say about these questions. I will be glad to answer any questions you might have after I have had the chance to hear you out' (Hessler 1992: 139). In short, orthodox methodologists hold that the key to extracting truthful replies lies in the reduction of bias from the interviewers, whose subjective beliefs and personalities might influence the way subjects respond, and the reduction of bias from respondents, whose subjective perceptions of the research aims, and of the interviewer, and of what is socially desirable and so on, might encourage them to lie, exaggerate, or otherwise conceal the truth.
Interviewing    121
For the orthodox social scientist, then, subjectivity is bias and ideally, the researcher should eliminate all traces of it in order to lay bare the objective truths behind it. Though it acknowledges the need for rapport, this approach demands that the interviewer performs the role, as far as possible, of a neutral instrument for extracting and recording a very specific and limited set of data. Any superfluous information provided by the respondent is ignored, his or her comments and questions, any detail of how the interviewer felt about the interview and so on, in fact, all traces of the interview as an interaction between two people are expunged from the final record. All this is in stark contrast to the approach recommended by qualitative and feminist methodologists.
QUALITATIVE APPROACHES TO ASKING QUESTIONS
Methods textbooks which take a more qualitative or an ethnographic approach to interviewing do not use the natural science language of 'variables', 'control', 'standardisation', and so on, but see the interview as an opportunity to delve and explore precisely those subjective meanings that positivists seek to strip away. Qualitative research is generally not so much concerned with obtaining accurate replies to closed ended questions, as with obtaining full and sincere responses to relatively open-ended enquiries. This is because most qualitative research is informed, to some degree, by the interpretative tradition described in Chapter 2. Where researchers wish to achieve some form of verstehen (interpretative understanding), both interviewer and interviewee will need to play very different roles to those set out above. The interviewee is not a research 'subject' to be controlled and systematically investigated by a 'scientist', but a reasoning, conscious human being to be engaged with. Hammersley and Atkinson (1989) observe that:
The main difference between the way in which ethnographers and survey interviewers ask questions is not, as is sometimes suggested, that one form of interviewing is 'structured' and the other is 'unstructured'. . . . The important distinction to be made is between standardized and reflexive interviewing.
(Hammersley and Atkinson 1989: 112-13)
Methodologists who take a more qualitative approach to interviewing argue that the interviewer must enter into an interaction with the interviewee and therefore needs to be prepared to respond flexibly to the interviewee as an individual, subjective being. Each interviewee and therefore each interview is accepted as different and individual,
122    Methods, sex and madness
regardless of whether a structured interview schedule is being followed or not. Some qualitative methodologists see interviews as spontaneous and unpredictable encounters, rejecting the idea of formulating questions and probes in advance. Glesne and Peshkin (1992), for example, state that listening is one of the most important acts performed by the interviewer:
At no time do you stop listening, because without the data your listening furnishes, you cannot make any of the decisions inherent in interviewing: . . . Has your question been answered, and is it time to move on? If so, move on to what question? Should you probe now or later? What form should your probe take? . . . The spontaneity and unpredictability of the interview exchange precludes planning your probes ahead of time; you must, accordingly, think and talk on your feet.
(1992: 76)
This view of the qualitative interview is perhaps overdrawn. Certainly, in qualitative research, interviewers are far less controlling than survey interviewers, and the interview is a far more flexible and responsive tool. But this does not mean interviewing calls for no pre-planning, or that interviewers just sit back passively and allow topics or issues to emerge. Qualitative interviews are generally focused rather than completely free-flowing and spontaneous, though the degree of fluidity and improvisation will depend on the stage of the research and how much good and reliable data has already been gathered. Qualitative researchers may conduct completely unstructured interviews in the early stages of research, allowing the interviewee to talk about whatever seems most important to him or her, but in the later stages of research, they may wish to confirm particular points, or to focus on particular issues or topics and will then stick more closely to pre-set questions or topic headings.
The significant difference between qualitative and orthodox interviews is that, whether structured or unstructured, the qualitative interview has more the character of a dialogue than a quest for simple 'yes' and 'no' responses. For this reason, qualitative methodologists do not express the ambivalence about rapport which is to be found in orthodox accounts of interviewing. Establishing a good rapport is a vital element of the interviewing process; 'rapport is tantamount to trust, and trust is the foundation for acquiring the fullest, most accurate disclosure a respondent is able to make' (Glesne and Peshkin 1992:79). Rather than trying to expunge the personality of the interviewer and to standardise interviews, this approach demands that interviewers should manage
Interviewing    123
their appearance, behaviour and self presentation in such a way as to build rapport and trust with each individual respondent. If rapport is not developed, the quality of information gathered during the course of the interview will suffer. Good rapport also allows the interviewer to keep the interview focused on the topics he or she wishes to hear about. An interviewee is more likely to allow you to change the subject, interrupt or 'steer' the conversation if he or she warms to you as a person. This means that where orthodox researchers value interviewers with merely a 'pleasant and businesslike manner', qualitative methodologists tend to ask for much more. Interviewers must be empathetic and committed:
When you are warm and caring, you promote rapport, you make yourself appealing to talk to, and, not least, you communicate to your respondents, 'I see you as a human being with interests, experience, and needs beyond those I tap for my own purposes.' ... In an effective interview, both researcher and respondent feel good, rewarded and satisfied by the process and the outcomes. The warm and caring researcher is on the way to achieving such effectiveness.
(Glesne and Peshkin 1992: 87)
Orthodox researchers see subjectivity as bias - something to be controlled or expunged. But in qualitative research, recognising the subjectivity of interviewer and interviewee is a key aspect of acquiring knowledge. Orthodox researchers assume that the people will tell the truth providing their responses are not contaminated by interaction between the interviewer and respondent as subjective beings. Qualitative methodologists, on the other hand, not only argue that interaction is necessary to develop a trusting relationship within which people will be prepared to disclose the truth, but also that a dialogue between the interviewer and interviewee as subjective beings is necessary to ensure that the interviewer has fully understood what is being disclosed. For the orthodox researcher, such a dialogue would prevent standardisation and require the interviewer to make subjective judgements and interpretations, and so would lead to bias. But for those who take a more interactive approach, double checking that the respondent understands the question and that the interviewer understands the reply is the only way to be sure of acquiring reliable, meaningful responses.
This concern with the interviewee as a subjective being does not require that the researcher abandons a commitment to objectivity, however. Qualitative methodologists, as much as those from the orthodox school, warn against the use of leading questions (see Glesne and Peshkin 1992), and stress that the interviewer should be conscious at all
124    Methods, sex and madness
times of how his or her line of questioning may be affecting the interviewee's responses. Hammersley and Atkinson also note that 'a useful tactic is to make the question "lead" in a direction opposite to that in which one expects the answer to lie and thus avoid the danger of simply and misleadingly confirming one's expectations' (1989: 115-16). Essentially, then, qualitative methodologists recognise the interview as a far more complex phenomenon than do their orthodox counterparts. This is also evident in relation to the issue of researcher control, which is seen as a paradox, rather than simply asserted to be a requirement for 'objective' fact gathering. Glesne and Peshkin (1992) point out that whilst non-hierarchical relationships between researcher and researched are sometimes possible:
In most instances . . . the researcher maintains a dominant role that reflects his or her definition of the inquiry purposes. As long as the purposes are his or her own, the researcher sustains a power imbalance that may or may not get redressed, depending on the researcher's opportunity for and commitment to reciprocity.
(1992: 82)
But though researchers are dominant in this respect, they are simultaneously submissive in as much as they 'cannot dictate the particulars of [the] interviewer-interviewee relationship' (Glesne and Peshkin 1992: 82). Getting access, eliciting continued co-operation throughout the course of the interview, how long the interview lasts, what is and is not discussed - all these crucial factors are within the gift of the interviewee, and thus outside the control of the interviewer. To this we could add that power relations external to the interview situation can further complicate the issue of control. A female researcher interviewing males may be in control in the sense that she is initiating the interview and using it to her own ends, but may simultaneously feel vulnerable because of her status as a woman. For example, O'Connell Davidson's research in the water industry involved her undertaking lengthy qualitative interviews, sometimes with solitary men in isolated geographical settings. For a woman to find herself in a secluded spot in the countryside, talking to a man she has never before met and unsure whether anyone else knows her whereabouts is extremely unnerving (see O'Connell Davidson 1991: 228). Likewise, for a female interviewer to be shut in an office with a lecherous senior manager is not conducive to a sense of full control.
The real point is that, for qualitative researchers, control is an ambiguous issue. Taking a non-directive approach to questioning and
Interviewing    125
allowing the interviewee, at least in part, to set the agenda, is not necessarily 'a methodological disaster that terminates useful data gathering' as Hessler (1992: 137), quoted earlier, would have it. The interviewer is always an active agent in the process of data collection. By listening to what the respondent chooses to say, rather than forcing him or her to answer simply a pre-set list of questions, the interviewer is obtaining more and better quality information, information which can help to shape the future course of the research (Hammersley and Atkinson 1989: 114). Interviewers are not losing their 'objectivity', becoming partial or imposing a particular world view on the respondent, rather they are using the interview as an opportunity to explore the subjective values, beliefs and thoughts of the individual respondent. Many of these same themes and issues have recently been taken up by feminist methodologists.
THE FEMINIST APPROACH TO ASKING QUESTIONS
In recent years, a number of feminist writers have begun to argue for a distinctive feminist methodology (see, for example, Fonow and Cook 1991, Stanley and Wise 1993). They challenge the claim that traditional social scientists produce value-free, objective knowledge, arguing that sexist value judgements explicitly and implicitly inform what people choose to study, how they go about investigating social phenomena and how they interpret their findings. The idea that social research is coloured by the values and preconceptions of the researcher is by no means peculiar to feminist writers, but the emphasis upon a specifically masculine bias in social research has been an important part of the feminist critique of methods. So far as interviewing is concerned, some feminists assert that the orthodox methodologisť s emphasis on control, hierarchy and the impersonal nature of scientific research reflects a masculine view of the world and of human relationships more generally. Oakley (1981) argues that the orthodox paradigm of the social research interview is characterised by the following features:
(a) its status as a mechanical instrument of data-collection; (b) its function as a specialised form of conversation in which one person asks the questions and another gives the answers; (c) its characterisation of interviewees as essentially passive individuals and (d) its reduction of interviewers to a question asking and rapport-promoting role.
(1981: 36-7)
126    Methods, sex and madness
This paradigm is rejected as morally indefensible, since it is controlling and exploitative, but Oakley also points to more practical reasons for ditching the advice of orthodox methods texts, noting that 'the goal of finding out about people through interviewing is best achieved when the relationship of interviewer and interviewee is non-hierarchical and when the interviewer is prepared to invest his or her own personal identity in the relationship' (1982: 41). Feminist methodologists therefore tend to recommend many of the techniques advocated by qualitative methodologists. Since they too are primarily concerned with obtaining full and sincere responses, the need for rapport and genuine interaction between interviewer and interviewee as subjective beings is stressed. The interviewer can answer the questions of the interviewee and enter into a genuine emotional relationship, possibly even a friendship with her research subjects. It is here that the similarities between feminist and qualitative approaches end. Qualitative methodologists make a strong distinction between conversations that are part of friendship and the research interview, noting that it is quite possible to achieve a rapport and learn a great deal from people that you do not like. They also recognise the complexity surrounding power and control in the interview relationship, whereas feminist methodologists often assume that where both interviewer and interviewee share the same gender socialisation and critical life-experiences, the interview can be a genuinely non-exploitative, non-hierarchical meeting of equals. Unlike qualitative methodologists, who recognise the need to retain some control over the focus and scope of the interview, many feminist methodologists seem to imply that researchers can and should just listen to women, allowing respondents to set their own agendas and focus on what is important to them. They insist that researchers should not try to control the research subject, but should instead attempt to work themselves into the mind of the social actor:
and see the world as he or she sees it; this is not controlling the actor and manipulating his or her behaviour but rather respecting people's integrity . . . [This produces] an honest accounting empathetic of people and events and a most illuminating perspective.
(Farganis 1989: 213)
This technique of simply letting women 'tell their stories' is supposed to enable the interviewer to enter the common-sense world of the women she studies and to document it. This non-exploitative, descriptive research is said to have an emancipatory potential. Since feminist methodologists explicitly seek to redress the gender imbalance in social
interviewing    127
research, they direct their advice towards women researchers interviewing other women, 'sisters' researching 'sisters', and argue that the knowledge such research produces can and should be returned to the community to empower the female subjects of feminist research.
Thus far, the differences between these three approaches to interviewing have been emphasised. Before turning to look at some examples of interviewing in the real world, it is worth noting that there are also some similarities. Orthodox, qualitative and feminist methodologists could all agree, for example, that an interview is not merely a conversation between two (or more) people. They would not agree on what precisely the rules and skills appropriate to interviewing are, but most would agree that interviewing is bound by rules of interaction and requires skills that are different to those employed in everyday conversation. Interviewers will not glean much information, for example, if they use the interview as an opportunity to hold forth on their own life experiences or political beliefs, whereas this might be acceptable as part of an ordinary conversation. Interviewers are unlikely to get the interviewee to co-operate with any type of interview if they are visibly bored, hostile or judgemental. The need for some degree of rapport is recognised by all methodologists, and furthermore, all advocate impartiality in as much as it is accepted that interviewers should not lead or manipulate respondents into saying what the interviewer wishes or expects to hear. The interviewing techniques used in a social services investigation into child sex abuse in Britain recently caused much controversy precisely because many people felt that interviewers put ideas into the children's heads and used leading questions to extract the answers they expected. The report of the inquiry commissioned to review the case describes one interview as follows:
[Tjhe interviewers immediately raised with WB the question of whether she knew why she was there. They indicated that people had been saying that she had been hurt on parts of her body, and asked WB if at any time she gathered with a group of people where she or others were touched or hurt in any way ... they had indicated to WB
that they believed the information which had been given to them___
When WB insisted that nothing had happened to her, the interviewers tried to reassure her that they would listen to her, but they continued to indicate their belief that she had been hurt.
(Report of the Inquiry into the Removal of Children from Orkney
1992: 185)
Orthodox, qualitative and feminist methodologists would all object to such an approach, recognising that if interviewers doggedly push people
128    Methods, sex and madness
into giving the answers or story they wish to hear, the data they produce is biased and worthless. In short, whether researchers use highly structured interview schedules or take a completely unstructured approach, and whether they are pursuing standardisation or reflexivity, the interview is not a chance encounter or an aimless chat. It is initiated by the researcher and takes place with a particular goal in mind, namely data collection. Reliable data will not be collected if the interviewer leads the interviewee, suggesting appropriate responses and refusing to acknowledge others. The following outline of interviewing methods in the real world vividly illustrate the issues which have been raised thus far.
INTERVIEWING KINSEY STYLE
It was noted in the previous chapter that most of the methodological criticism of Kinsey's work has centred upon his sampling techniques. Even recent commentators comment favourably upon his interviewing methods. Shipman (1988), for example, describes Kinsey's interviews as follows:
Complete confidentiality, absolute privacy during interviewing and no suggestions of right or wrong behaviour were the guides to rapport. Kinsey himself carried out 7,000 of the interviews lasting an hour to an hour and a half. This labour of love was conducted deadpan; friendly, but never with any expression of surprise or disapproval. The questions were asked as directly as possible to avoid interaction. The interviewer looked squarely at the subject and moved inexorably from factual background to intimate detail. . . . Questions were used to check others, husbands were checked against wives, reinterviewing after eighteen months was employed. This study is acknowledged as a classic.
(1988: 84)
Kinsey's own descriptions of his interviewing methods certainly match this portrait of the impartial scientist, rigorously pursuing the truth, but accounts of the interviews offered by his associates Johnson, Pomeroy and Gebhard paint a rather different picture. Let us return to the central problem facing the orthodox interview researcher: how do you get people to give you truthful and accurate responses? Pomeroy states that there were only three possible ways that Kinsey's subjects could not tell the truth: they could exaggerate their sexual experience, they could remember events incorrectly, or they could deny and otherwise conceal their sexual experience. Only the last way of not telling the truth was
Interviewing    129
believed by Kinsey and his associates to represent a real problem for his research:
Exaggeration was almost impossible with the system we used for asking questions rapidly and in detail. People who tried reported little success. Not remembering accurately could be dealt with statistically; the errors one person might make were offset by errors another made in the opposite direction. Covering up was the most serious problem.
(Pomeroy 1972: 120)
In other words, Kinsey and his associates assumed that the real problem was going to be getting people to admit the range and extent of their sexual experience, not that they might exaggerate it. Having decided in advance that 'covering up' would be the main barrier to truth, Kinsey held that respondents must be asked questions in such a way as to ensure that they felt free to admit anything. He therefore insisted that respondents should never be asked whether they had experienced sexual intercourse, cunnilingus, or whatever, but always asked when they had first done it. This was intended not only to show the interviewee how relaxed and non-judgemental the interviewer was, but also to make it more difficult for them to 'cover up' certain activities. As Pomeroy (1972) explains, it meant that "The subject who might want to deny an experience had a heavier burden placed on him, and since he knew from the way the question was asked that it would not surprise us if he had done it, there seemed little reason to deny it' (1972: 112, emphasis added). This may have been all very well when they wanted to find out whether someone had ever masturbated an animal, but placing the burden of denial upon the interviewee could well have had the effect of exaggerating the incidence of more mundane forms of sexual activity. To ask a 17-year-old boy, for example, when he first had full penetrative sex, rather than if he has ever had full penetrative sex, or to ask a married man when he first, rather than if, he had extra-marital affairs, could be described as leading. The difference between asking a question in such a way as to suggest that a given response would not surprise the interviewer, and asking a question in such a way as to suggest that the interviewer expects that given response is rather fine.
To assess how likely Kinsey's respondents were to have been frank and open in the course of the interviews, it is also important to ask questions about the interviewers themselves.
The attributes of the interviewer would be identified by all three of the different approaches to interviewing outlined above as an important factor affecting how willing interviewees would be to disclose the truth.
130    Methods, sex and madness
Qualitative and feminist methodologists, for example, would probably recommend that interviewers be matched to respondents in terms of gender, and possibly also in terms of age and ethnicity, in order to set respondents at ease and encourage a good rapport to develop. Kinsey took a rather different view. Pomeroy explains that:
It was suggested to us that we ought to have women interviewers to interview women, and Negro interviewers for blacks. By that logic, Kinsey pointed out, we would have to have prostitutes for prostitutes, drug addicts for drug addicts and so on. The qualities of the interviewer, not his sex, race or personal history, were the important variables.
(1972: 102)
If sex and 'race' were really unimportant to Kinsey, it seems curious that all of his interviewers just happened to be white Anglo-Saxon Protestant (WASP) males, and that he even expressed concern about hiring anyone 'with a Jewish name because he thought some WASP interviewees might object' (Pomeroy 1972: 103). We are then hurriedly reassured that 'there was nothing bigoted* about Kinsey's refusal to employ women, Black people or people with Jewish names, it was simply that he believed 'only WASPs... could interview everybody' (1972: 102). This highlights a very real problem with various methodologists' assertion that the characteristics of the interviewer can either encourage or discourage people from participating in the research. It is actually very hard to disentangle the researcher's own prejudices from his or her fears about potential interviewees' prejudices. Were people really more likely to respond to a WASP interviewer than to a Jewish interviewer, or was it simply that Kinsey preferred to hire WASPs than to hire Jewish people?
Kinsey's assumption that women would tell the truth about their sexual lives to male interviewers certainly now appears as dated. How much effect this exclusive use of male interviewers had upon Kinsey's findings is, of course, impossible to determine, but it is interesting to note that Russell's research into incestuous abuse, which did match interviewers and respondents in terms of gender and ethnicity, found that 12 per cent of a random sample of women reported having been abused, whereas only around 2 to 3 per cent of Kinsey's female sample disclosed such experiences (see Russell 1986: 64-5). (This discrepancy may also reflect the different sampling methods employed. It may be that women who had suffered sexual abuse in childhood were less likely to volunteer to take part in sex research.) Similarly, the extent to which the exclusive use of WASP interviewers affected respondent's willingness to disclose the truth cannot be accurately determined, but the
Interviewing    131
description of interviewing practice provided by Pomeroy is extremely offensive and suggests that Kinsey's faith in the WASP's ability to 'interview everybody' was misplaced. Moving immediately on from a discussion about how rapport was achieved with child respondents, Pomeroy explains:
It was particularly important that we know the sexual viewpoint of the cultures from which our subjects came. Kinsey illustrated this point with the case of an older Negro male who at first had been wary and evasive in his answers. From the fact that he listed a number of minor jobs when asked about his occupation and seemed reluctant to go into any of them, [Kinsey] deduced that he might have been active in the underworld, so he began to follow up by asking the man whether he had ever been married. He denied it, at which Kinsey resorted to the vernacular and inquired if he had ever 'lived common law'. The man admitted he had, and that it had first happened when he was fourteen.
'How old was the woman?' [Kinsey] asked.
'Thirty-five,* he admitted, smiling.
Kinsey showed no surprise. 'She was a hustler, wasn't she?' he said flatly. At this, the subject's eyes opened wide, he smiled in a friendly way for the first time, and said, 'Well, sir, since you appear to know something about these things, I'll tell you straight.'
(Pomeroy 1972: 115-16)
Why was Kinsey so quick to assume that this man was 'active in the underworld' or that the woman he mentioned was a 'hustler'? Why did he not consider the possibility that the man was exaggerating the age difference, or that the smile and wide-eyed 'Well, sir' could have been a straightforward mockery of Kinsey's rather transparent line of thought? This really leads into another major problem with Kinsey's interviewing methods, namely the question of how Kinsey and his associates decided whether or not to accept an interviewee's replies as true. The extract quoted above gives an example of Kinsey suggesting to a respondent that he had lived with a 'hustler' at the age of 14 and accepting his failure to deny this as positive confirmation. On other occasions, however, the researchers did not accept the respondent's initial responses as gospel. Gebhard and Johnson (1979: 20) comment that when a respondent reported unusually high frequencies of sexual intercourse, for example, it was viewed with suspicion and the interviewer would ask a series of additional questions to double check. This generally led to the respondent modifying the original claim. However,
132    Methods, sex and madness
'we only utilized this technique in cases where our suspicions were aroused and, consequently, exaggerations which fell within the range of probability passed uncorrected* (1979: 21). Similarly, respondents were subjected to the third degree when the researchers thought that their answers were wrong or incomplete. The same question would be rephrased and asked again. Pomeroy notes that if he or Kinsey thought that an interviewee was lying, they would pretend to have misunderstood his reply and say things like: 'Yes, I know you have never done that, but how old were you the first time that you did it?' This approach was particularly fruitful with people termed by Pomeroy as 'of low mentality* (1972: 113). The questions placed the burden of denial upon the subject, but if the interviewer did not believe their denials, pressure was applied. Pomeroy explains how, if they were convinced a subject was lying, he and Kinsey would challenge them:
It became necessary to say, with firmness, even vehemence, and yet always with kindness, 'Look, I don't give a damn what you've done, but if you don't tell me the straight of it, it's better that we stop this history right here. Now, how old were you the first time that this or that happened?' Surprisingly, in not a single case did a person refuse to continue.
(Pomeroy 1972: 127)
Considered in the light of research into social conformity and the way in which people tend to comply with researchers who are perceived as authority figures, this makes Kinsey's figures on levels of sexual activity look even more suspect. If subjects did admit to being homosexual, or to having had animal contacts or whatever after being pressured in this way, their 'confession' would be recorded with no mention of what had gone on between interviewer and respondent in order to elicit the information. Pomeroy notes that To make it easy for subjects to correct answers, we ignored contradictions, accepting the correction as though it were a first reply' (1972: 113). The same approach was taken with the follow up interviews. If, after the interview was over, they began to suspect that it contained falsities they would return to the subject and 'demand that he correct the record' (Pomeroy 1972: 113). This technique of demanding 'corrections', then accepting and recording them as if they were a first reply means that Kinsey's interviewing methods rather falls between two stools. He was using orthodox coding procedures (that is, recording only the limited and specific response and no detail of the interaction between interviewer
Interviewing    133
and interviewee that led to this response) combined with highly unorthodox procedures for extracting those responses.
More importantly though, we need to ask what it was that suddenly made Kinsey and his associates recognise the 'falsity' of an interview? What was it that made them able to tell whether a subject was lying or covering up, and so decide to rephrase questions, interrogate further or threaten to terminate the interview? Why was it that the African-American respondent's claim to have lived with a 35-year-old woman at the age of 14 was accepted and recorded without doubt or further question, whilst other claims made by other respondents were disbelieved and challenged? Could it have been that Kinsey chose only to accept as true those replies which fitted his pre-existing theories and his pre-existing stereotypes about what kind of sexual behaviour a given 'type' of person might indulge in? All this underlines the point that the social research interview does not simply raise the question of how to get people to disclose the truth, but also the more intractable methodological question of how researchers can and do decide which answers to accept and record as true. This problem emerges equally forcefully in relation to the interviewing techniques adopted by Sigmund Freud.
INTERVIEWS WITH SIGMUND FREUD
Any methodological critique of Freud's interviewing techniques is complicated by the fact that his interviews were not undertaken simply for research purposes, but were also intended to be of therapeutic value to the interviewee. He himself might therefore explain certain of his interviewing practices in relation to his role as a doctor, rather than as a researcher. Freud may, for example, have insisted upon complete control of the interviews because he held this to be appropriate to the doctor-patient relationship, rather than because he imagined this to be the best way of extracting the truth in a researcher-researched relationship. Likewise, Freud held that the interviewee/patient's condition precluded any genuinely intimate or equal relationship with him as interviewer/doctor. Though Freud's interviewing methods undoubtedly match the masculine paradigm described and criticised by feminist methodologists, it is more difficult to condemn them as straightforwardly exploitative and morally indefensible, since psychotherapy was ultimately intended to benefit the patient. However, as well as being designed to serve therapeutic purposes, Freud's interviews served to produce the data upon which he developed his psychoanalytic theories,
134    Methods, sex and madness
and here we are solely concerned with his interviewing techniques as a method of data collection. Freud claimed to have undertaken a scientific study of the unconscious, and it is therefore worth examining his interview practice in the light of the issues raised in the above sections.
Freud first graduated as doctor of medicine, and undertook research into the clinical use of cocaine. He then spent a year in Paris with Charcot, studying nervous diseases, particularly 'hysteria'. On his return to Vienna, Freud became interested in a method first pioneered by Josef Breuer, a consultant who argued that hysteria was the product of a trauma which had been forgotten by the patient (Bocock 1986: 1). The treatment consisted of using hypnosis to get the patient to recall the forgotten event and live through the appropriate emotional response to it. For Freud, the significance of hypnosis was that it revealed the existence of active parts of the mind that are not generally discernible either to the individual or the on-looker. He pointed out that in a hypnotic trance, people remember details about their lives that they cannot normally recall. Moreover, though people do not remember consciously what has been said to them during hypnosis, they will later act on suggestions made to them by the hypnotist. From this he concluded that there exists a part of the mind which is inaccessible to individuals at conscious level, yet still influences what they do, how they feel and so on. Freud called this hidden part of the mind the unconscious, and his aim was to scientifically explore its structure and content. However, whilst hypnosis had given Freud proof of the existence of this unconscious mind, he saw it as an imperfect research instrument. It was erratic and irregular. Sometimes it worked, at other times it did not. Some subjects were open to hypnosis, others were not. Freud explains:
I soon came to dislike hypnosis, for it was a temperamental and, one might almost say, a mystical ally. ... I set about working with patients in their normal state. At first, I must confess, this seemed a senseless and hopeless undertaking. I was set the task of learning from the patient something that í did not know and that he did not know himself. How could one hope to elicit it?
(Freud 1974: 47)
How indeed? Freud wanted a method which would get people to tell the truth about their unconscious mind, something which, by definition, they did not consciously know anything about. The solution came to Freud when he realised that people were actually capable of dragging memories from the unconscious into the conscious mind without help of
Interviewing     135
hypnosis. This he deduced from the fact that although people who had been hypnotised would initially say that they remembered nothing of what had been said during hypnosis, it would eventually be recalled if they were put under sufficient pressure. He therefore decided to use the same technique without hypnosis:
When I reached a point with [my patients] at which they maintained they knew nothing more, I assured them that they did know it all the same, and that they had only to say it; and I ventured to declare that the right memory would occur to them at the moment at which I laid my hand on their forehead.
(1974: 47)
Freud found this technique worked. Patients who at first told him they could not remember certain events or scenes from their childhood would produce memories when he laid his hand upon their forehead. As a form of scientific investigation this is problematic. If an interviewer asks 'Did you ever witness your parents in flagranteV and refuses to accept the respondent's claim not to remember any such thing, saying 'Yes, you can remember, tell me about it,' it would certainly convey the impression that the interviewer both wanted and expected the respondent to produce a suitable memory. If the interviewer then said, 'When I lay my hand on your forehead, the right memory will occur to you,' an enormous pressure to comply would be exerted. To fail to come up with a memory in the face of such explicit expectations would appear uncooperative. Moreover, Freud's interviewees were typically unhappy or disturbed, probably feeling vulnerable having labelled themselves, or been labelled by their family as 'hysterical', and it is therefore likely that they were in a highly suggestible state. On top of this, as a doctor, Freud was an authority figure and the power relationship between him and his patients would presumably have added to this pressure to produce appropriate memories. A recent biography of Graham Greene notes that both Greene and his cousin Ave were sent to a Jungian therapist for treatment. This involved arriving at the therapist's office at 11 a.m. each morning and recounting their dreams of the previous night. Both Greene and his cousin often found it impossible to recall their dreams, and therefore, as Ave remarked, the two of them 'used to concoct dreams' for the therapist to analyse. Greene began all his fictitious dreams with a pig (Sherry 1989:96).
Although Freud was keen to stress the 'scientific' nature of his research, there can be little doubt that his interaction with his patients affected what they told him. His line of questioning was leading in the extreme. He
136    Methods, sex and madness
assumed that all neuroses originated in sublimated or repressed infantile wishes, interpreted everything the patient said and every dream they recounted through reference to childhood experiences, and then fed these interpretations back to the patient. Certainly orthodox methodologists would argue that such interviewing techniques would contaminate the responses provided by patients, for if respondents have a clear idea of the theory or hypothesis under investigation and the interviewer's concerns, they are likely to select out responses to please the interviewer - in this case to produce memories, dreams and ideas which fitted with Freud's obsession with repressed wishes and desires. Neither would Freud's approach be acceptable to qualitative methodologists. Freud was not engaging with his patients as subjective beings, but was quite relentlessly imposing his own agenda and beliefs upon them. It seems unlikely that such techniques are conducive to getting people to tell you the whole truth and nothing but the truth about their thoughts, dreams or experiences. Freud's theoretical preconceptions informed the questions he asked, and the way in which he asked his questions must surely have coloured the responses he elicited. His theoretical preconceptions also affected whether or not he accepted a patient's answers as true. Freud's abandonment of the seduction theory provides a clear example of how critical this issue is to research findings.
When Freud first started practising, he was visited by many 'hysterics', who displayed symptoms such as a nervous tic or cough, depression, or suicidal feelings and many of these people claimed that they had been sexually abused in their childhood. Following Breuer, Freud initially saw these traumatic childhood experiences as the source of the adult hysteria. In other words, he began by believing what his patients told him. He accepted their accounts of rape and molestation as true descriptions of events that had really taken place. The hysterical symptoms they developed in later life were a response to real events that had taken place. But as we saw in Chapter 1, these ideas were not well received by the medical establishment and Freud quickly abandoned his seduction theory. He came to see his former willingness to accept the word of his patients at face value as 'naive'. He now thought that his female patients were describing to him their unconscious wishes and desires, rather than giving him accurate and reliable accounts of events. When women told him that their fathers had raped or molested them, Freud now believed that they were mistaking their wish for sexual contact with their father for reality. Nothing actually happened, but the girls longed for it to happen. They then came to see their longing as unacceptable and had to repress it, and it was this, not a real experience of abuse, which lay at the heart of their neurosis. What they told him in interviews was not truth but fantasy.
Interviewing    137
The question of whether a patient's account of childhood sexual traumas were accepted as true or rejected as fantasy was vital to the whole future development of Freud's research, theories and therapeutic practice. It meant the difference between searching for the origins of neuroses in the real world, or locating them in the internal world of the subject's unconscious mind (Masson 1984). The shift away from accepting patient's statements as true also means that the methodological foundations of Freud's theories are extremely shaky. He asked people questions, and concluded that the truth was the precise opposite of what they told him. He gathered together empirical evidence to the effect that many of his patients had suffered childhood sexual abuse, and used it to develop a theory which said they had not. Few methodologists could be happy with such an approach to data collection or analysis. The problem is well illustrated by one of Freud's case studies, that of a young woman, referred to as Dora. The following is a summary of Porter's (1989) excellent outline of her case.
Freud found Dora to have various 'hysterical' symptoms, including a nervous cough, general debility, migraine, and a disposition to flirt with suicide. She told Freud that an old family friend, Herr K., had kissed her when she was about 14-years-old, and sexually propositioned her three years later. When Dora said she found the man's advances disgusting, Freud took this to be a hysterical symptom, stating that a 'healthy' girl would find it pleasant and exciting to be kissed by a close friend of her father's. Freud claimed that in reality, Dora desired Herr K., but this desire conflicted with her Oedipal longing for her father, and she therefore had to deny and repress her true feelings for Herr K. When Freud put this to Dora, she denied it. So Freud explained that when she resisted and rejected his line of analysis, she was really confirming that it was true:
To an objective observer like himself, such a denial really meant confirmation. Patients said 'No' in their consciousness. But, Freud explained, 'there is no such thing as an unconscious No'. . . . Likewise, Freud assured her, when a patient, denying an interpretation, says 'I didn't think like that,' the real meaning of the phrase is 'Yes, I was unconscious of that'. All this is, Freud assures his readers, an 'entirely trustworthy form of confirmation*.
(Porter 1989: 115)
Again, we have to ask questions about when and why researchers accept that people are telling the truth, when and why they reject their informant's accounts. When Freud first lectured on his seduction theory, the orthodox medical profession rejected it because it was based on the
138    Methods, sex and madness
word of 'hysterical' women. It seems that ultimately Freud came to share the view that such women are unreliable informants, especially when they make claims against those more powerful and better respected than themselves (i.e., against men). A researcher who accepts the word of a relatively weak and powerless group at face value risks vilification and professional isolation. The social pressure to make research findings acceptable, to record only that which fits with received ideas and commonly accepted views of the world can be immense.
If we concentrate solely upon Freud's interviewing techniques, asking whether they were likely to get people to talk openly and sincerely, and whether Freud was likely to judge accurately when the truth had been told, it is difficult to avoid concluding that Freud's own theoretical framework both affected what people told him, and coloured his judgements as to what was true and untrue. Equally, it is important to recognise that research does not take place in a social and political vacuum, but in the context of a particular set of normative and moral values. Freud undertook his work at a time when women were considered to be so intellectually and emotionally inferior that they did not even enjoy full rights of citizenship. The fact that, in this context, he jettisoned a theory which rested on the uncorroborated evidence of 'hysterical' women is perhaps unsurprising, but his willingness to shift blame away from the patriarchs and on to their victims is also a very clear example of the kind of sexism which feminist methodologists argue has distorted social and psychological research. The following section looks at the interview research Diana Scully recently conducted with convicted rapists. Again, it highlights the problem which interviewers face in trying to get respondents to give accurate, full and frank responses, but it also allows us to explore some of the limitations of the approach to interviewing recommended by feminist methodologists.
INTERVIEWING WITHOUT SYMPATHY
Diana Scully's Understanding Sexual Violence (1990) is based upon interviews, conducted in prisons by herself and a colleague, with 114 convicted rapists. Scully is a feminist and fiercely critical of the sexist bias in traditional social research, but she points out that, important as it is to do research into women's lives, feminist researchers must do more than, simply describe the experiences of women:
I continue to be concerned that feminist scholars are neglecting another . . . area of critical work on men's world. . . . While not
interviewing    139
diminishing the continuing responsibility to illuminate women's subordinate condition, the debunking of patriarchy is not accomplished by focusing exclusively on the lives and experiences of women.
(Scully 1990: 3)
This raises important issues for advocates of a feminist methodology. If such writers wish to distinguish feminist methodology from orthodox methods by an insistence that feminist researchers be genuinely non-exploitative and truly care about their research subjects, they will pretty much preclude any feminist studies of men who oppress and/or brutalise women. Since orthodox methodology is held to be inadequate because it cannot accommodate the experience of one half of the population, it would be odd for feminists to champion a method which could only be applied to the other half. It also highlights the danger of a method which seeks only to 'give voice' to research subjects, rather than critically explore and analyse their worldviews. Feminists may be happy to empathetically describe the experiences of their female subjects, but presumably would not wish to use the same method with male subjects.
Scully argues that it is essential to understand the men and the ideologies that oppress women, and illustrates the point through reference to rape. Feminist researchers have undertaken important research into the experiences of women who have survived rape; they have looked at the psychological and legal problems such women experience, and have challenged the many horrific but widely accepted myths about women and rape. While such work is vital to the struggle to change things like the way that rape victims are treated by the police and legal system, it cannot hope to address the question of why some men are sexually violent. This area has been largely left to male researchers to investigate, and Scully argues that male researchers have tended to explain rape in terms of individual pathology, using a disease model, rather than linking sexual violence to social beliefs and attitudes. Furthermore, much of this research has done little more than reproduce and reinforce existing social attitudes towards rape, effectively blaming women for men's sexual violence (see Scully 1990: 33-60).
Scully therefore argues the case for feminist research with rapists and adopts a socio-cultural framework which considers rape not in terms of individual pathology, but as an extension of normal masculinity. She observes that far from being abberant, expressions of power and domination are socially prescribed and rewarded to men in North American society. The aim of her study was to discover what men who rape gain from their sexually violent behaviour and to explore their more general
140    Methods, sex and madness
attitudes towards masculinity, femininity and sexual violence in an attempt to uncover links between their sexual violence and broader social ideas about women, masculinity and violence. This involved undertaking highly structured, but essentially qualitative interviews with convicted rapists, as well as with a control group of felons convicted of non-sexual offences. The first and most striking feature of her interviews is simply their length:
The interviews were long - for rapists, 89 pages divided into three parts: Part 1 consisting of a complete background history including childhood, family, religious, marital, education, employment, sexual and criminal; Part 2 consisting of a series of scales measuring attitudes toward women, masculinity, interpersonal violence, and rape; and Part 3 consisting of 40 pages of open-ended questions about the rape and the victim. Those in the contrast group of other felons were given only Part 1 and 2 of the interview.
(Scully 1990: 14)
Clearly getting people to even co-operate with an interview such as this is no easy task, as Scully notes 'the success of this research hinged on the ability to develop a good working relationship within a very brief span of time' (1990: 14). Without trust and rapport, no one is likely to disclose much about their marital and sexual history, far less details about their own criminal behaviour. Given the subject matter of the interviews, it would have been extremely difficult for Scully to follow the advice of orthodox methodologists to the letter. She explains that:
I. . . found that it was impossible to adhere to a rigid sequence for questions. Quite simply, no matter how much probing was done, the men would not talk about certain things until they were ready and felt comfortable. So although all of the questions were asked of every man, the interviews, rather than being uniform, were all slightly different, depending on the needs and readiness of the interviewee.
(1990: 14-15, emphasis added)
Yet it would also have been hard for Scully to feel genuinely 'warm and caring* about the interviewees, as some qualitative methodologists recommend, or to follow certain feminist methodologists' advice about respecting the subject's integrity and taking a non-hierarchical approach to the interview. Scully does actually manage to show the reader the world through the eyes of the rapists she interviewed. She shows how these men rationalised and justified their actions, how they made themselves believe that their horrific acts were acceptable, necessary or
Interviewing    141
inevitable, and she is able to do this because during the course of the interviews she managed to get them to open up and talk in detail about their attitudes and their crimes, about how and why they chose to rape. But she did not enter into their worlds through a process of empathetic identification, she did not build trust and rapport upon a foundation of genuine warmth or sympathy. Instead she manipulated these men, encouraging them to see her in a particular way so that they would be willing to disclose information that they may perhaps not have disclosed had they known how she truly felt about them:
The type of information sought in this research required a supportive, non judgmental neutral facade - one that I did not always genuinely feel. Frankly, some of the men were personally repulsive.. . . Additionally the stories they told were horrible and a few of the men were not overly co-operative. Indeed, some of the interviews required immense effort to remain neutral. But the fact is that no one tells his or her secrets to a visibly hostile and disapproving person.
(Scully 1990: 18)
Few people would be filled by a sense of moral outrage at the idea of manipulating convicted rapists in this way, but it is worth noting that such an approach can generate other ethical dilemmas. Scully goes on to observe that her non-judgemental facade may have had unintended consequences; 'This was especially problematic with the men . . . who did not define their sexually violent behaviour as rape. I worried that some of these men might interpret neutrality as a signal of agreement or approval' (1990: 18-19). Her dilemma was intensified by the fact that if she interacted with these men as a person, rather than as a 'neutral' researcher, and disagreed with or challenged their views, they could have told other potential interviewees and 'who would volunteer to get shot down by the researcher?' (Scully 1990: 19). Scully's interviews also raise other ethical problems. She argues that in order to get convicted rapists to talk truthfully about their attitudes and behaviour, it was necessary to assure them that whatever they said to her would have no bearing on their future parole or release, and to guarantee them complete confidentiality (except if interviewees confided plans for future illegal acts outside the prison). She justifies this guarantee as follows:
In the case of past criminal behaviour, confidentiality is justified because past acts do not pose a current threat, and since the individual is already in prison, no one is placed in danger from the continued activity. Likewise, the confidentiality of details about ongoing illegal
142    Methods, sex and madness
activities that occur within prisons, involving things like drugs or sexual behaviour, can be defended. Since staff are generally aware that these activities exist to some degree in all prisons, informing about a specific act would contribute little to what is already known or to a solution.
(1990: 23)
This formulation seems inadequate. What would have happened if a rapist, nearing the end of a relatively short sentence for one rape, had confessed to her a series of undetected rapes and murders? Surely under such circumstances the researcher would have a moral obligation to pass this information on to the authorities and so break the promise of confidentiality. The case for confidentiality as regards illegal sexual behaviour within the prison seems equally weak. If a man confessed to her that he was raping other men in prison, why should she protect him? It could equally well be argued that a researcher is under no obligation to inform the police if he or she extracts information about undetected rapes outside prison - after all, the police are generally aware that rapes occur, and informing about a specific act would likewise 'contribute little to what is already known or to a solution'. However, though all ethical issues are not resolved by the formulation quoted above, Scully was undoubtedly correct to assume that without a guarantee of confidentiality, her interviewees would have been less likely to disclose information, and that if they had believed that their chances of parole could be affected by what they told her, it would have been a positive incentive to He or conceal the truth. In short, to encourage these men to tell the whole truth in the interviews, Scully had to build and maintain trust and rapport which involved concealing her true feelings about them, refraining from challenging or making negative comments about their assertions, and assuring them complete confidentiality. AH of these things could, for one reason or another, be judged by a purist to be unethical, and yet without them it is unlikely that Scully would have gathered much useful data.
Scully also had to face the problem of how to decide whether or not her interviewees were telling the truth. Unlike many researchers, she confronts this issue directly in a section of her chapter on methodology. (Indeed, Scully should be more generally congratulated for her unusually detailed and thoughtful discussion of methodology.) So far as questions about the actual rapes for which the men were convicted, Scully had access to an independent source of verification in the form of pre-sentence reports, written by court workers at the time of conviction
Interviewing    143
which provide information on both the offender's, victim's and police's versions of the details of the crime (Scully 1990: 31). Each interviewee was asked a list of 30 factual questions about his background and about the crime, and then, with the interviewee's permission, his responses were checked against these pre-sentence reports 'to establish the validity of the interview' (Scully 1990: 26). The availability of these records was of enormous value to the study. They allowed Scully to distinguish between three distinct types of rapists. First, there were those men who admitted the rape (admitters) and whose version of events in the interview broadly corresponded with that provided in the presentence report. Scully notes, however, that whilst they did not actually tell lies:
they did systematically understate the amount of force and violence they used. . . . Admitters also did not volunteer information about especially brutal or offensive aspects of their crimes. For example, a particularly anguished young man tearfully recounted the details of his rape, including the age of his 70 year old victim. His self-disgust was further clarified when the validity check revealed what he neglected to mention - that the victim had been his grandmother and that she suffered a heart attack as a result of the rape.
(1990: 27)
Second, there were those who admitted having had sexual contact with their victims, but denied that they had raped them (deniers). The information they gave Scully in the interviews differed markedly from the victim's and police's versions of what had taken place. These men 'seemed genuinely to believe that their actions were not rape despite the admission, in some cases, that a weapon had been used* (1990: 27). Finally, there were 34 men who denied any contact at all with their victims, 'instead, they said that they themselves were the victims of mistaken identity, or that they had not raped the victims but had committed other crimes against them, such as robbery' (1990: 28).
Unlike Kinsey and Freud, Scully did not contradict her interviewees or challenge them to correct their answers if she suspected they were lying, but recorded their statements as they stood. The men who claimed to have no knowledge of the rapes they were convicted of were not questioned further about the rapes - Scully collected only background information and attitudinal data from them. So far as the men who admitted having had sexual contact but denied rape were concerned, Scully did not believe that what they told her was a true account of events, but she did believe they were sincerely describing what they saw
144    Methods, sex and madness
as the truth. She suggests that deniers were not setting out to deceive her as a researcher, but rather were honestly describing to her how they deceived themselves. Accepting what they told her as a truthful portrayal of their self-deception actually formed an important part of her analysis. She argues that 'denials can also be taken at face value, and the content analysed as a statement on the cultural learning and socially derived perspective of sexually violent men' (1990: 28). It might be argued that Freud approached his patient's denials in a similar way, but a critical difference is that Freud had no way of knowing whether the denials were self-deception other than through reference to his own theories, whereas Scully was able to corroborate her suspicions through reference to documentary evidence compiled by others. Clearly, she could only independently verify issues covered by the pre-sentence report. Responses to other questions and attitudinal data could not be checked, and it is possible that these men expressed attitudes that they did not sincerely hold. But, as will be argued below, there is a sense in which interview research always relies on the good faith of the interviewee. Methodologists can argue the toss about which interviewing techniques are most likely to encourage people to be truthful, but the assumption that having volunteered or agreed to take part in research, the interviewee will not systematically and deliberately lie in response to every question remains an act of faith.
CONCLUSIONS
Earlier, we noted that all methodologists can agree upon the fact that interviews differ from ordinary conversations, and that it is wrong to lead or manipulate respondents into providing the answers that the researcher (for theoretical, political or moral reasons) wishes to hear. What orthodox, qualitative and feminist methodologists differ on is the role of the interviewer, the nature of the interviewer-interviewee relationship, and how exactly to go about extracting reliable data. Advice on these matters varies not simply because methodologists draw on very different philosophical traditions, but also because interviewing can be used in such a vast range of different types of research. The role of the interview in survey research is very different, and usually far more limited, than the interview is in ethnographic research. No one who wished to grasp the meanings that give form and content to social processes in an alien culture, for example, would set out to administer a pre-designed, standardised set of questions and no one who simply wanted to know whether people preferred butter or margarine would put
Interviewing    145
enormous effort into establishing an excellent rapport and enter into a lengthy, in-depth, unstructured interview. The role of the interview in research which sets out to test a particular hypothesis is likewise different from that of the interview in theory-constructing research, and it is different in research which aims to produce an almost literary, descriptive account of, say, six prostitute's lives (see Jaget 1980) than in research which aims to systematically document the response of 30 women clerical workers to the introduction of new office technology (see O'Connell Davidson 1994). In other words, the scope of the interview (and therefore many of its characteristics) is powerfully affected by the researchers objectives and by other aspects of the research design. But whatever the aim and scope of the interview, it is always a social encounter, and this fact gives rise to the paradoxes which cannot be fully resolved by philosophical or methodological dogma. Let us spell these out before concluding.
Theorising about the virtues of various different interviewing techniques and actually conducting interviews are two rather different matters and, in part, this has to do with the fact that researchers cannot control for individual differences between respondents. When you talk to people who have conducted social research interviews or read accounts of the interviewing techniques adopted in particular research projects, you find that, no matter how committed the researcher was originally to a particular theoretical model of interviewing, the practice of interviewing diverged from the theoretical ideal to some degree. Interviewers who believe in the central importance of standardisation will come across people who are just plain awkward. Almost inevitably, there will be some people whose personalities are such that they cannot be put through a standard interview and who will manage to get even the best trained interviewer to answer some query or explain some item in more detail. Qualitative interviewers will also come up against people who they cannot interview in the recommended manner. Some people are just not expansive, and no matter how much effort is put into building trust and rapport they will still answer every question with a monosyllable and shrug their shoulders or look away in embarrassment when they are probed further. Even feminist interviewers interviewing women and committed to a non-directive, non-hierarchical approach can be confronted by 'sisters' who are so loquacious or so determined to discuss trivia that they are forced to adopt a more directive and controlling style.
Most methodologists observe that the success of the interview, in terms of extracting reliable data, relies to some degree on the personal
146    Methods, sex and madness
characteristics and social identity of the interviewer. Some will insist that the blander the interviewer the better, others will argue that interviewers and respondents should be 'matched' in terms of gender and/or 'racialised' identity, some will call for interviewers to be caring, sharing, 'feely' people. Against this, we would argue that it is impossible to prescribe the 'best* or 'necessary* characteristics of an interviewer, other than to say that the interviewer should be able to exploit their personal characteristics to full advantage. Interviewers can and do employ what might appear to be a handicap, such as a stutter, to their advantage. Embarrassment sometimes encourages interviewees to keep talking and to say more than they otherwise would. In some situations, interviewers can and do exploit the negative stereotypes that certain people hold of them to their advantage. A young women interviewing relatively powerful male respondents, senior managers for example, can live up to their expectations by 'acting dumb' which can encourage them to disclose more information to her than they would to an older male interviewer, whom they might assume would know how to use the information against them. Kinsey's insistence on WASP male interviewers underlines the moral and political dangers of making the case for a certain 'type' of interviewer, while Scully's work demonstrates the fact that differences in gender are not necessarily a barrier to obtaining full and detailed information, even about topics of an extremely sensitive nature.
It is also important to recognise that there is no method or technique which actually forces people to provide truthful and accurate responses, far less full and sincere ones, and other than strapping your subjects to a lie-detector (which is, in any case, an unreliable instrument) there is no way of knowing with certainty whether or not they have told you the truth. Ultimately, deciding that an answer is true and complete is a subjective judgement on the part of the interviewer. This judgement can be informed by experience (both as a member of society and as a professional) and/or by more ephemeral qualities such as intuition. Certain skills, such as reading body language and facial movements, or being able to quickly cross-reference one statement with other previous statements, are obviously also useful. However, whatever this process of deciding whether a response is true or false may be, it is not 'scientific' in the natural science sense of the word. As a consumer of research, you therefore need to assess how likely the researcher was to have made good judgements about whether their informants were telling the truth. You need to ask yourself how likely the researcher was to be swayed by their own theoretical, moral and normative preconceptions. No matter
Interviewing     147
which general approach to interviewing is taken (orthodox, qualitative or feminist), these preconceptions can lead interviewers to refuse to accept disconfirming statements from interviewees and to be too willing to accept confirming statements at face value, without further probing. All this points to the following conclusion. There are certain interview practices which we can condemn universally. These include any techniques which lead or manipulate the respondent into saying that which the researcher wishes to hear, or which prevent the respondent from stating that which he or she wishes to state. But beyond this, it is not possible to lay down blanket rules and procedures to cover every conceivable social research interview that every single interviewer could successfully follow, any more than it would be possible to write a blueprint for how to behave in any other social encounter that every single person could use effectively. Any handbook which pretends otherwise is necessarily downplaying the complexity of the social interaction that interviewing involves, and the range of uses to which interviews can be put in social research.