Interviews

CHAPTER 21

Interviews are a widely used instrument for data collection. This chapter sets out a range of key issues to be considered in planning, conducting and reporting interviews, including:

   conceptions of the interview

   purposes of the interview

   types of interview

   planning interview-based research procedures

   group interviewing

   interviewing children

   interviewing minority and marginalized people

   focus groups

   non-directive, focused, problem-centred and in-depth interviews

   telephone interviewing

   ethical issues in interviewing

This chapter indicates different kinds of interview, and argues for ‘fitness for purpose’ to be addressed in deciding which kind of interview and interview questions to employ.

21.1 Introduction

The use of the interview in research marks a move away from seeing human subjects as simply manipulable and data as somehow external to individuals, and towards regarding knowledge as generated between humans, often through conversations (Kvale, 1996: 11). Regarding an interview, as Kvale (1996: 14) remarks, as an inter-view, an interchange of views between two or more people on a topic of mutual interest, sees the centrality of human interaction for knowledge production, and emphasizes the social situatedness of research data. As we suggested in Chapter 5, knowledge should be seen as constructed between participants, generating data rather than capta (Laing, 1967: 53). As such, the interview is not exclusively either subjective or objective, it is intersubjective (Laing, 1967: 66). Interviews enable participants – be they interviewers or interviewees – to discuss their interpretations of the world in which they live, and to express how they regard situations from their own point of view. In these senses the interview is not simply concerned with collecting data about life: it is part of life itself, its human embeddeness is inescapable.

The interview is a flexible tool for data collection, enabling multi-sensory channels to be used: verbal, non-verbal, spoken and heard. The order of the interview may be controlled whilst still giving space for spontaneity, and the interviewer can press not only for complete answers but for responses about complex and deep issues. In short, the interview is a powerful implement for researchers. On the other hand the researcher using interviews has to be aware that they are expensive in time, they are open to interviewer bias, they may be inconvenient for respondents, issues of interviewee fatigue may hamper the interview, and anonymity may be difficult. We explore these several issues in this chapter.

An interview is not an ordinary, everyday conversation (Dyer, 1995: 56–8). For example, in contrast to an everyday conversation, it has a specific purpose, it is often question-based, with the questions being asked by the interviewer; the interviewer alone may express ignorance (and not the interviewee), and the responses must be as explicit and often as detailed as possible. The interview is a constructed and usually a specifically planned event rather than naturally occurring situation, and this renders it different from an everyday conversation; therefore the researcher has an obligation to set up, and abide by, the different ‘rules of the game’ in an interview.

21.2 Conceptions of the interview

Kitwood (1977) lucidly contrasts three conceptions of an interview. The first conception is that of a potential means of pure information transfer. He explains that:

if the interviewer does his job well (establishes rapport, asks questions in an acceptable manner, etc.), and if the respondent is sincere and well-motivated, accurate data may be obtained. Of course all kinds of bias are liable to creep in, but with skill these can largely be eliminated. In its fullest expression, this view accords closely with that of the psychometricians, who apparently believe that there is a relatively permanent, consistent, ‘core’ to the personality, about which a person will give information under certain conditions. Such features as lying, or the tendency to give a socially desirable response, are to be eliminated where possible.

(Kitwood, 1977)

This conception of the interview appears to be widely held.

A second conception of the interview is that of a transaction which inevitably has bias, that needs to be recognized and controlled. According to this viewpoint, Kitwood (1977) explains that ‘each participant in an interview will define the situation in a particular way. This fact can be best handled by building controls into the research design, for example by having a range of interviewers with different biases.’ The interview is best understood in terms of a theory of motivation which recognizes a range of non-rational factors governing human behaviour, like emotions, unconscious needs and interpersonal influences. Kitwood points out that both these views of the interview regard the inherent features of interpersonal transactions as if they were ‘potential obstacles to sound research, and therefore to be removed, controlled, or at least harnessed in some way’.

The third conception of the interview sees it as an encounter necessarily sharing many of the features of everyday life (see for example, Box 21.1). Kitwood (1977) suggests that what is required, according to this view, is not a technique for dealing with bias, but a theory of everyday life that takes account of the relevant features of interviews. These may include role-playing, stereotyping, perception and understanding. As Walford (2001: 90) remarks, ‘interviewers and interviewees co-construct the interview’. The interview is a social encounter, not simply a site for information exchange, and researchers would be well advised to keep this in the forefront of their minds when conducting an interview.

One of the strongest advocates of this latter viewpoint is Cicourel (1964) who lists five of the unavoidable features of the interview situation that would normally be regarded as problematic.

1   There are many factors which inevitably differ from one interview to another, such as mutual trust, social distance and the interviewer’s control.

2   The respondent may well feel uneasy and adopt avoidance tactics if the questioning is too deep.

3   Both interviewer and respondent are bound to hold back part of what it is in their power to state.

4   Many of the meanings which are clear to one will be relatively opaque to the other, even when the intention is genuine communication.

5   It is impossible, just as in everyday life, to bring every aspect of the encounter within rational control.

The message that proponents of this view would express is that no matter how hard an interviewer may try to be systematic and objective, the constraints of everyday life will be a part of whatever interpersonal transactions she initiates. Kitwood concludes:

The solution is to have as explicit a theory as possible to take the various factors into account. For those who hold this view, there are not good interviews and bad in the conventional sense. There are simply social encounters; goodness and badness are predicates applicable, rather, to the theories within which the phenomena are explained.

(Kitwood, 1977)

BOX 21.1    ATTRIBUTES OF ETHNOGRAPHERS AS INTERVIEWERS

Trust: There would have to be a relationship between the interviewer and interviewee that transcended the research, that promoted a bond of friendship, a feeling of togetherness and joint pursuit of a common mission rising above personal egos.

Curiosity: There would have to be a desire to know, to learn people’s views and perceptions of the facts, to hear their stories, discover their feelings. This is the motive force, and it has to be a burning one, that drives researchers to tackle and overcome the many difficulties involved in setting up and conducting successful interviews.

Naturalness: As with observation one endeavours to be unobtrusive in order to witness events as they are, untainted by one’s presence and actions, so in interviews the aim is to secure what is within the minds of interviewees, uncoloured and unaffected by the interviewer.

Source: Adapted from Woods, 1986

Indeed Barker and Johnson (1998: 230) argue that the interview is a particular medium for enacting or displaying people’s knowledge of cultural forms, as questions, far from being neutral, are couched in the cultural repertoires of all participants, indicating how people make sense of their social world and of each other.1

21.3 Purposes of the interview

The purposes of the interview in the wider context of life are many and varied, for example:

   to evaluate or assess a person in some respect;

   to select or promote an employee;

   to effect therapeutic change, as in the psychiatric interview;

   to test or develop hypotheses;

   to gather data, as in surveys or experimental situations;

   to sample respondents’ opinions, as in door-step interviews.

Although in each of these situations the respective roles of the interviewer and interviewee may vary and the motives for taking part may differ, a common denominator is the transaction that takes place between seeking information on the part of one and supplying information on the part of the other.

As a distinctive research technique, the interview may serve three purposes. First, it may be used as the principal means of gathering information having direct bearing on the research objectives. As Tuckman (1972) describes it, ‘By providing access to what is “inside a person’s head”, [it] makes it possible to measure what a person knows (knowledge or information), what a person likes or dislikes (values and preferences), and what a person thinks (attitudes and beliefs).’ Second, it may be used to test hypotheses or to suggest new ones; or as an explanatory device to help identify variables and relationships. And third, the interview may be used in conjunction with other methods in a research undertaking. In this connection, Kerlinger (1970) suggests that it might be used to follow up unexpected results, for example, or to validate other methods, or to go deeper into the motivations of respondents and their reasons for responding as they do.

As our interests lie primarily in reviewing research methods and techniques, we will subsequently limit ourselves to the use of the interview as a specific research tool. Interviews in this sense range from the formal interview in which set questions are asked and the answers recorded on a standardized schedule through less formal interviews in which the interviewer is free to modify the sequence of questions, change the wording, explain them or add to them to the completely informal interview where the interviewer may have a number of key issues which she raises in conversational style instead of having a set questionnaire. Beyond this point is located the non-directive interview in which the interviewer takes on a subordinate role.

The research interview has been defined as ‘a two-person conversation initiated by the interviewer for the specific purpose of obtaining research-relevant information, and focused by him on content specified by research objectives of systematic description, prediction, or explanation’ (Cannell and Kahn, 1968). It is an unusual method in that it involves the gathering of data through direct verbal interaction between individuals. In this sense it differs from the questionnaire where the respondent is required to record in some way her responses to set questions.

As the interview has some things in common with the self-administered questionnaire, it is frequently compared with it. Each has advantages over the other in certain respects. The advantages of the questionnaire, for instance, are: it tends to be more reliable because it is anonymous; it encourages greater honesty; it is more economical than the interview in terms of time and money and there is the possibility that it may be mailed. Its disadvantages, on the other hand, are: there is often too low a percentage of returns; the interviewer is able to answer questions concerning both the purpose of the interview and any misunderstandings experienced by the interviewee, for it sometimes happens in the case of the latter that the same questions have different meanings for different people; if only closed items are used, the questionnaire will be subject to the weaknesses already discussed; if only open items are used, respondents may be unwilling to write their answers for one reason or another; questionnaires present problems to people of limited literacy; and an interview can be conducted at an appropriate speed whereas questionnaires are often filled in hurriedly.

By way of interest, we illustrate the relative merits of the interview and the questionnaire in Table 21.1. It has been pointed out that the direct interaction of the interview is the source of both its advantages and disadvantages as a research technique (Borg, 1963). One advantage, for example, is that it allows for greater depth than is the case with other methods of data collection. A disadvantage, on the other hand, is that it is prone to subjectivity and bias on the part of the interviewer.

TABLE 21.1 SUMMARY OF RELATIVE MERITS OF INTERVIEW VERSUS QUESTIONNAIR

Consideration	Interview	Questionnaire
1 Personal need to collect data	Requires interviewers	Requires a secretary
2 Major expense	Payment to interviewers	Postage and printing
3 Opportunities for response-keying (personalization)	Extensive	Limited
4 Opportunities for asking	Extensive	Limited
5 Opportunities for probing	Possible	Difficult
6 Relative magnitude of data reduction	Great (because of coding)	Mainly limited to rastering
7 Typically, the number of respondents who can be reached	Limited	Extensive
8 Rate of return	Good	Poor
9 Sources of error	Interviewer, instrument, coding, sample	Limited to instrument and sample
10 Overall reliability	Quite limited	Fair
11 Emphasis on writing skill	Limited	Extensive

Source: Tuckman, 1972

Oppenheim (1992: 81–2) suggests that interviews have a higher response rate than questionnaires because respondents become more involved and, hence, motivated; they enable more to be said about the research than is usually mentioned in a covering letter to a questionnaire, and they are better than questionnaires for handling more difficult and open-ended questions.

21.4 Types of interview

The number of types of interview given is frequently a function of the sources one reads! For example LeCompte and Preissle (1993) give six types: (i) standardized interviews; (ii) in-depth interviews; (iii) ethnographic interviews; (iv) elite interviews; (v) life history interviews; (vi) focus groups. Bogdan and Biklen (1992) add to this: (vii) semi-structured interviews; (viii) group interviews. Lincoln and Guba (1985) add: (ix) structured interviews; and Oppenheim (1992: 65) adds to this: (x) exploratory interviews. Patton (1980: 206) outlines four types: (xi) informal conversational interviews; (xii) interview guide approaches; (xiii) standardized open-ended interviews; (xiv) closed quantitative interviews. Patton sets these out clearly as shown in Table 21.2.

How is the researcher to comprehend the range of these various types? Kvale (1996: 126–7) sets the several forms of interview along a series of continua, arguing that interviews differ in the openness of their purpose, their degree of structure, the extent to which they are exploratory or hypothesis-testing, whether they seek description or interpretation, or whether they are largely cognitive-focused or emotion-focused. A major difference appears to lie in the degree of structure in the interview, which, itself, reflects the purposes of the interview, for example, to generate numbers of respondents’ feelings about a given issue or to indicate unique, alternative feelings about a particular matter. Lincoln and Guba (1985: 269) suggest that the structured interview is useful when the researcher is aware of what she does not know and therefore is in a position to frame questions that will supply the knowledge required, whereas the unstructured interview is useful when the researcher is not aware of what she does not know, and therefore relies on the respondents to tell her!

The issue here is of ‘fitness for purpose’; the more one wishes to gain comparable data – across people, across sites – the more standardized and quantitative one’s interview tends to become; the more one wishes to acquire unique, non-standardized, personalized information about how individuals view the world, the more one veers towards qualitative, open-ended, unstructured interviewing. Indeed, this is true not simply of interviews but of their written counterpart – questionnaires. Oppenheim (1992: 86) indicates that standardization should refer to stimulus equivalence, i.e. that every respondent should understand the interview question in the same way, rather than replicating the exact wording, as some respondents might have difficulty with, or interpret very differently, and perhaps irrelevantly, particular questions. (He also adds that as soon as the wording of a question is altered, however minimally, it becomes, in effect, a different question!)

Oppenheim (1992: 65) suggests that exploratory interviews are designed to be essentially heuristic and seek to develop hypotheses rather than to collect facts and numbers. He notes that these frequently cover emotionally loaded topics and, hence, require skill on the part of the interviewer to handle the interview situation, enabling respondents to talk freely and emotionally and to have candour, richness, depth, authenticity, honesty about their experiences.

TABLE 21.2 STRENGTHS AND WEAKNESSES OF DIFFERENT TYPES OF INTERVIEW

Type of interview	Characteristics	Strengths	Weaknesses
1 Informal conversational interview	Questions emerge from the immediate context and are asked in the natural course of things; there is no predetermination of question topics or wording.	Increases the salience and relevance of questions; interviews are built on and emerge from observations; the interview can be matched to individuals and circumstances.	Different information collected from different people with different questions. Less systematic and comprehensive if certain questions don’t arise ‘naturally’. Data organization and analysis can be quite difficult.
2 Interview guide approach	Topics and issues to be covered are specified in advance, in outline form; interviewer decides sequence and working of questions in the course of the interview.	The outline increases the comprehensiveness of the data and makes data collection somewhat systematic for each respondent. Logical gaps in data can be anticipated and closed. Interviews remain fairly conversational and situational.	Important and salient topics may be inadvertently omitted. Interviewer flexibility in sequencing and wording questions can result in substantially different responses, thus reducing the comparability of responses.
3 Standardized open-ended interviews	The exact wording and sequence of questions are determined in advance. All interviewees are asked the same basic questions in the same order.	Respondents answer the same questions, thus increasing comparability of responses; data are complete for each person on the topics addressed in the interview. Reduces interviewer effects and bias when several interviewers are used. Permits decision makers to see and review the instrumentation used in the evaluation. Facilitates organization and analysis of the data.	Little flexibility in relating the interview to particular individuals and circumstances; standardized wording of questions may constrain and limit naturalness and relevance of questions and answers.
4 Closed quantitative interviews	Questions and response categories are determined in advance. Responses are fixed; respondent chooses from among these fixed responses.	Data analysis is simple; responses can be directly compared and easily aggregated; many short questions can be asked in a short time.	Respondents must fit their experiences and feelings into the researcher’s categories; may be perceived as impersonal, irrelevant and mechanistic. Can distort what respondents really mean or experienced by so completely limiting their response choices.

Source: Patton, 1980: 206

Morrison (1993: 34–6) sets out five continua of different ways of conceptualizing interviews. At one end of the first continuum are numbers, statistics, objective facts, quantitative data; at the other end are transcripts of conversations, comments, subjective accounts, essentially word-based qualitative data.

At one end of the second continuum are closed questions, multiple choice questions where respondents have to select from a given, predetermined range of responses that particular response which most accurately represents what they wish to have recorded for them; at the other end of the continuum are much more open-ended questions which do not require the selection from a given range of responses – respondents can answer the questions in their own way and in their own words, i.e. the research is responsive to participants’ own frames of reference and response.

At one end of the third continuum is a desire to measure responses, to compare one set of responses with another, to correlate responses, to see how many people said this, how many rated a particular item as such-and-such; at the other end of the continuum is a desire to capture the uniqueness of a particular situation, person or programme – what makes it different from others, i.e. to record the quality of a situation or response.

At one end of the fourth continuum is a desire for formality and the precision of numbers and prescribed categories of response where the researcher knows in advance what is being sought; at the other end is a more responsive, informal intent where what is being sought is more uncertain and indeterminate – we only know what we are looking for when we have found it! The researcher goes into the situation and responds to what emerges.

At one end of the fifth continuum is the attempt to find regularities – of response, opinions, etc. – in order to begin to make generalizations from the data, to describe what is happening; at the other end is the attempt to portray and catch uniqueness, the quality of a response, the complexity of a situation, to understand why respondents say what they say, and all of this in their own terms.

One can cluster the sets of poles of the five continua thus:

Quantitative approaches	Qualitative approaches
numbers	words
predetermined, given	open-ended, responsive
measuring	capturing uniqueness
short-term, intermittent	long-term, continuous
comparing	capturing particularity
correlating	valuing quality
frequencies	individuality
formality	informality
looking at	looking for
regularities	uniqueness
description	explanation
objective facts	subjective facts
describing	interpreting
looking in from the outside	looking from the inside
structured	unstructured
statistical	ethnographic, illuminative

The left-hand column is much more formal and preplanned to a high level of detail, whilst the right-hand column is far less formal and the fine detail only emerges once the researcher is in situ. Interviews in the left-hand column are front-loaded, that is, they require all the categories and multiple choice questions to be worked out in advance. This usually requires a pilot to try out the material and refine it. Once the detail of this planning is completed the analysis of the data is relatively straightforward because the categories for analysing the data have been worked out in advance, hence data analysis is rapid.

The right-hand column is much more end-loaded, that is, it is quicker to commence and gather data because the categories do not have to be worked out in advance, they emerge once the data have been collected. However, in order to discover the issues that emerge and to organize the data presentation, the analysis of the data takes considerably longer.

Kvale (1996: 30) sets out key characteristics of qualitative research interviews, namely that they should:

   engage, understand and interpret the key feature of the lifeworlds of the participants;

   use natural language to gather and understand qualitative knowledge;

   be able to reveal and explore the nuanced descriptions of the lifeworlds of the participants;

   elicit descriptions of specific situations and actions, rather than generalities;

   adopt a deliberate openness to new data and phenomena, rather than being too pre-structured;

   focus on specific ideas and themes, i.e. have direction, but avoid being too tightly structured;

   accept the ambiguity and contradictions of situations where they occur in participants, if this is a fair reflection of the ambiguous and contradictory situation in which they find themselves;

   accept that the interview may provoke new insights and changes in the participants themselves;

   regard interviews as an interpersonal encounter, with all that this entails;

   be a positive and enriching experience for all participants.

There are four main kinds of interview that we discuss here that may be used specifically as research tools: (i) the structured interview; (ii) the unstructured interview; (iii) the non-directive interview; and (iv) the focused interview. The structured interview is one in which the content and procedures are organized in advance. This means that the sequence and wording of the questions are determined by means of a schedule and the interviewer is left little freedom to make modifications. Where some leeway is granted her, it too is specified in advance. It is therefore characterized by being a closed situation. In contrast to it in this respect, the unstructured interview is an open situation, having greater flexibility and freedom. As Kerlinger (1970) notes, although the research purposes govern the questions asked, their content, sequence and wording are entirely in the hands of the interviewer. This does not mean, however, that the unstructured interview is a more casual affair, for in its own way it also has to be carefully planned.

The non-directive interview as a research technique derives from the therapeutic or psychiatric interview. The principal features of it are the minimal direction or control exhibited by the interviewer and the freedom the respondent has to express her subjective feelings as fully and as spontaneously as she chooses or is able. Moser and Kalton (1977: 297) argue that respondents should be encouraged to talk about the subject under investigation (e.g. themselves) and to be free to guide the interview, with few set questions or pre-figured frameworks. The interviewer should prompt and probe, pressing for clarity and elucidation, rephrasing and summarizing where necessary and checking for confirmation of this, particularly if the issues are complex or vague.

The need to introduce rather more interviewer control into the non-directive situation led to the development of the focused interview. The distinctive feature of this type is that it focuses on a respondent’s subjective responses to a known situation in which she has been involved and which has been analysed by the interviewer prior to the interview. She is thereby able to use the data from the interview to substantiate or reject previously formulated hypotheses. As Merton and Kendall explain,

In the usual depth interview, one can urge informants to reminisce on their experiences. In the focused interview, however, the interviewer can, when expedient, play a more active role: he can introduce more explicit verbal cues to the stimulus pattern or even represent it. In either case this usually activates a concrete report of responses by informants.

(Merton and Kendall, 1946: 542)

We shall be examining both the non-directive interview and the focused interview in more detail later in the chapter.

21.5 Planning interview-based research procedures

Kvale (1996: 88) sets out seven stages of an interview investigation that can be used to plan this type of research: thematizing; designing; interviewing; transcribing; analysing; verifying; and reporting. We use these to structure our comments here about the planning of interview-based research.

Thematizing

The preliminary stage of an interview study will be the point where the purpose of the research is decided. It may begin by outlining the theoretical basis of the study, its broad aims, its practical value and the reasons why the interview approach was chosen. There may then follow the translation of the general goals of the research into more detailed and specific objectives. This is the most important step, for only careful formulation of objectives at this point will eventually produce the right kind of data necessary for satisfactory answers to the research problem.

Designing

There follows the preparation of the interview schedule itself. This involves translating the research objectives into the questions that will make up the main body of the schedule. This needs to be done in such a way that the questions adequately reflect what it is the researcher is trying to find out. It is quite usual to begin this task by writing down the variables to be dealt with in the study. As one commentator says, ‘The first step in constructing interview questions is to specify your variables by name. Your variables are what you are trying to measure. They tell you where to begin’ (Tuckman, 1972).

Before the actual interview items are prepared, it is desirable to give some thought to the question format and the response mode. The choice of question format, for instance, depends on a consideration of one or more of the following factors:

   the objectives of the interview;

   the nature of the subject matter;

   whether the interviewer is dealing in facts, opinions or attitudes;

   whether specificity or depth is sought;

   the respondent’s level of education;

   the kind of information she can be expected to have;

   whether or not her thought needs to be structured;

   some assessment of her motivational level;

   the extent of the interviewer’s own insight into the respondent’s situation;

   the kind of relationship the interviewer can expect to develop with the respondent.

Having given prior thought to these matters, the researcher is in a position to decide whether to use open and/or closed questions, direct and/or indirect questions, specific and/or non-specific questions and so on.

Construction of schedules

Three kinds of items are used in the construction of schedules used in research interviews (see Kerlinger, 1970). First, ‘fixed-alternative’ items allow the respondent to choose from two or more alternatives. The most frequently used is the dichotomous item which offers two alternatives only: ‘yes/no’ or ‘agree/disagree’, for instance. Sometimes a third alternative such as ‘undecided’ or ‘don’t know’ is also offered.

Example: Do you feel it is against the interests of a school to have to make public its examination results?

Yes

No

Don’t know

Kerlinger has identified the chief advantages and disadvantages of fixed-alternative items. They have, for example, the advantage of achieving greater uniformity of measurement and therefore greater reliability; of making the respondents answer in a manner fitting the response category; and of being more easily coded.

Disadvantages include their superficiality; the possibility of irritating respondents who find none of the alternatives suitable; and the possibility of forcing responses that are inappropriate, either because the alternative chosen conceals ignorance on the part of the respondent or because she may choose an alternative that does not accurately represent the true facts. These weaknesses can be overcome, however, if the items are written with care, mixed with open-ended ones, and used in conjunction with probes on the part of the interviewer.

Second, ‘open-ended items’ have been succinctly defined by Kerlinger (1970) as ‘those that supply a frame of reference for respondents’ answers, but put a minimum of restraint on the answers and their expression’. Other than the subject of the question, which is determined by the nature of the problem under investigation, there are no other restrictions on either the content or the manner of the interviewee’s reply.

Example: What kind of television programmes do you most prefer to watch?

Open-ended questions have a number of advantages: they are flexible; they allow the interviewer to probe so that she may go into more depth if she chooses, or to clear up any misunderstandings; they enable the interviewer to test the limits of the respondent’s knowledge; they encourage cooperation and help establish rapport; and they allow the interviewer to make a truer assessment of what the respondent really believes. Open-ended situations can also result in unexpected or unanticipated answers which may suggest hitherto unthought-of relationships or hypotheses. A particular kind of open-ended question is the ‘funnel’ to which reference has been made earlier. This starts, the reader will recall, with a broad question or statement and then narrows down to more specific ones. Kerlinger (1970) quotes an example from the study by Sears et al. (1957):

All babies cry, of course. Some mothers feel that if you pick up a baby every time it cries, you will spoil it. Others think you should never let a baby cry for very long. How do you feel about this? What did you do about it? How about the middle of the night?

(Sears et al, 1957: 157)

Third, the ‘scale’ is, as we have already seen, a set of verbal items to each of which the interviewee responds by indicating degrees of agreement or disagreement. The individual’s response is thus located on a scale of fixed alternatives. The use of this technique along with open-ended questions is a comparatively recent development and means that scale scores can be checked against data elicited by the open-ended questions.

Example: Attendance at school after the age of 14 should be voluntary:

Strongly agree	Agree	Undecided
Disagree	Strongly disagree

It is possible to use one of a number of scales in this context: attitude scales, rank-order scales, rating scales and so on. We touch upon this subject again subsequently.

In devising questions for the interview, attention has to be given to (Arksey and Knight, 1999: 93–5):

   the vocabulary to be used (keeping it simple);

   the avoidance of prejudicial language;

   the avoidance of ambiguity and imprecision;

   leading questions (a decision has to be taken whether it is justified to use them);

   the avoidance of double-barrelled questions (asking more than one point at a time);

   questions that make assumptions (e.g. do you go to work in your car?);

   hypothetical or speculative questions;

   sensitive or personal questions (whether to ask or avoid them);

   assuming that the respondent has the required knowledge/information;

   recall (how easy it will be for the respondent to recall memories).

Question formats

We now look at the kinds of questions and modes of response associated with interviewing. First, the matter of question format: how is a question to be phrased or organized? (see Wilson, 1996). Tuckman (1972) has listed four such formats that an interviewer may draw upon. Questions may, for example, take a direct or indirect form. Thus an interviewer could ask a teacher whether she likes teaching: this would be a direct question. Or else she could adopt an indirect approach by asking for the respondent’s views on education in general and the ways schools function. From the answers proffered, the interviewer could make inferences about the teacher’s opinions concerning her own job. Tuckman suggests that by making the purpose of questions less obvious, the indirect approach is more likely to produce frank and open responses.

There are also those kinds of questions which deal with either a general or specific issue. To ask a child what she thought of the teaching methods of the staff as a whole would be a general or non-specific question. To ask her what she thought of her teacher as a teacher would be a specific question. There is also the sequence of questions designated the funnel in which the movement is from the general and non-specific to the more specific. Tuckman (1972) comments, ‘Specific questions, like direct ones, may cause a respondent to become cautious or guarded and give less-than-honest answers. Nonspecific questions may lead circuitously to the desired information but with less alarm by the respondents.’

A further distinction is that between questions inviting factual answers and those inviting opinions. To ask a person what political party he supports would be a factual question. To ask her what she thinks of the current government’s foreign policy would be an opinion question. Both fact and opinion questions can yield less than the truth, however: the former do not always produce factual answers; nor do the latter necessarily elicit honest opinions. In both instances, inaccuracy and bias may be minimized by careful structuring of the questions.

There are several ways of categorizing questions, for example (Spradley, 1979; Patton, 1980):

   descriptive questions;

   experience questions;

   behaviour questions;

   knowledge questions;

   construct-forming questions;

   contrast questions (asking respondents to contrast one thing with another);

   feeling questions;

   sensory questions;

   background questions;

   demographic questions.

These concern the substance of the question. Kvale (1996: 133–5) adds to these what might be termed the process questions, i.e. questions that:

   introduce a topic or interview;

   follow up on a topic or idea;

   probe for further information or response;

   ask respondents to specify and provide examples;

   directly ask for information;

   indirectly ask for information;

   interpret respondents’ replies.

We may also note that an interviewee may be presented with either a question or a statement. In the case of the latter she will be asked for her response to it in one form or another.

Example question: Do you think homework should be compulsory for all children between 11 and 16?

Example statement: Homework should be compulsory for all children between 11 and 16 years old.

Agree Disagree Don’t know

Stylianou (2008) discusses the ‘interview control question’. In experimental and survey designs, variables are often controlled, i.e. held ‘constant and unvarying so that one can see the true effects of other variables’ after the effects of others have been neutralized (controlled out), i.e. what effects remain after all the other variables have been controlled out (Morrison, 2009: 65). Morrison continues:

For example, we might be interested in examining the effects of gender on earnings, but these earnings are also affected by, for example, education level, type of work, qualifications, ethnicity, previous job experience, social behaviour and personality factors. If we wish to see the effects of gender then we have to control these other factors. This means that, in the case of gender for example, we would have to keep the males and females matched on the other variables, so that we could gain a true picture of the effects of gender rather than, say, education or job experience.

(Morrison, 2009: 65)

He suggests that controlling for the effects of other variables can be undertaken, inter alia, through randomization and random allocation (see Chapter 16), and isolation and control of independent variables of those other than the one(s) in which the researcher is interested (e.g. holding them constant). Controlling for the effects of other variables enables the researcher to see the true effect(s) of a single independent variable in which she or he is interested, i.e. what is left after other variables have been controlled out of the situation. Stylianou (2008) suggests that the same can be done in interviews, i.e. by isolating and controlling out the effects of other variables the researcher can see the true effect of a particular variable in which she or he is interested, i.e. when the effects of others have been removed. Interview control questions are a form of a probe (discussed below).

Let us give an example of an interview control question in an imaginary interview concerning a parent who expresses a negative attitude towards mixed ability classes in a primary school:

1   Interviewer. Why are you not in favour of mixed ability classes in the school?

2   Respondent: The less able students will slow down the more able students in the class, and the teacher will have to work very hard to keep up with the wide range of different abilities in the class.

3   Interviewer. But we know that many more capable students slow down anyway, for two reasons: firstly, if they finish work quickly then they are given more work to do, and they want an easy life, and secondly, many of the more able students don’t want to stand out as being exceptional in their class, so they slow down. And next, the teacher has to work hard anyway, as she has a range of tasks to do as part of her daily work. In fact the teacher can have a classroom assistant to work with students of different abilities.

4   Respondent: But having a classroom assistant still doesn’t make sure that all the students get their fair share of the teacher’s attention – only the less able and more able children get the extra attention from the classroom assistant.

5   Interviewer: But that’s the case anyway, as not all the students get the same amount of attention by the teacher, regardless of their abilities, as some students prefer to work quietly on their own without the teacher. Students have to work by themselves anyway, for example, in their mathematics lessons only they can do the work and the teacher cannot do it for them. And, anyway, it’s important for students to learn to work by themselves; isn’t that a good thing?

6   Respondent: But some students aren’t good at working by themselves and they may need the teacher’s help at critical moments, and having so many mixed abilities will prevent the teacher from being there at critical moments.

7   Interviewer: But the teacher will be there to help them at critical moments anyway, that’s part of their job, and they are trained to recognize critical moments. Teachers have to be present at critical moments in a student’s thinking, prompting them to take the next step in their thinking or learning.

8   Respondent: But some students will want to have an easy life, so they won’t let the teacher know that they need help or prompting, and they will ask their friend to help them.

9   Interviewer: But students do that anyway, as they often help each other, surely that’s a good thing, to work collaboratively and help each other, and students learn well from each other.

10   Respondent: Look, I just don’t want my child to have to work with less able children, and that’s all.

In the example, the interviewer is carefully stripping away the possible causes of the parent’s negative attitude: (a) slowing down the more capable students (lines 2 and 3); (b) students having fair access to adult help (lines 4 and 5); (c) students having to learn by themselves (lines 5 and 6); (d) the presence of teachers at critical moments (lines 6 and 7); (e) students working with friends (lines 8 and 9). The interviewer is raising alternative applications of each of the possible causes, i.e.

   there are reasons other than the one given here as to why the teacher has to work hard anyway (i.e. not only a matter of having the more and less able students in the same class), and why the more able students may slow down their rate of learning, not only the presence of less able students;

   there are reasons other than the one given here as to why having a classroom assistant will not help to solve the problem of students’ access to the teacher’s attention (i.e. there are other things that a classroom assistant has to do);

   there are reasons other than the one given here as to why students work by themselves, not only a matter of having or not having the teacher’s attention;

   there are reasons other than the one given here as to why the teacher may be present at critical moments;

   there are reasons other than the one given here as to why students work together.

The interviewer is finding that the reasons that the respondent gives for objecting to mixed ability classes are not substantial, as these reasons operate in other contexts as well, and not solely mixed ability contexts, and so they have to be controlled out: teacher working hard; access to teacher’s attention; students working on their own; teacher’s non-presence at critical moments; students working collaboratively.

In line 10, the respondent, having had a range of variables controlled (neutralized) here (a) to (e), becomes exasperated and ends that part of the interview. The researcher might conclude here that the parent is simply prejudiced, other key variables having been removed (controlled) from the reasoning. (Indeed Stylianou (2008: 244) suggests that this kind of probing is useful for studying attitudes and prejudice.) Here the interview control question assumes that the variables are dichotomous (e.g. the presence or absence of a variable are the only options); however, within that limitation, the interview control question is useful for identifying causal factors in an interviewees’ responses.

Response modes

If there are varied ways of asking questions, it follows there will be several ways in which they may be answered. It is to the different response modes that we now turn. In all, Tuckman (1972) lists eight such modes.

The first of these is the ‘unstructured response’. This allows the respondent to give her answer in whatever way she chooses.

Example: Why did you not go to university?

A ‘structured response’, by contrast, would limit her in some way.

Example: Can you give me two reasons for not going to university?

Although the interviewer has little control over the unstructured response, it does ensure that the respondent has the freedom to give her own answer as fully as she chooses rather than being constrained in some way by the nature of the question. The chief disadvantage of the unstructured response concerns the matter of quantification. Data yielded in the unstructured response are more difficult to code and quantify than data in the structured response.

A ‘fill-in response’ mode requires the respondent to supply rather than choose a response, though the response is often limited to a word or phrase.

Example: What is your present occupation?

or:

How long have you lived at your present address?

The differences between the fill-in response and the unstructured response is one of degree.

A ‘tabular response’ is similar to a fill-in response though more structured. It may demand words, figures or phrases, for example:

University	Subject	Degree	Dates
			From	To

It is thus a convenient and short-hand way of recording complex information.

A ‘scaled response’ is one structured by means of a series of gradations. The respondent is required to record her response to a given statement by selecting from a number of alternatives.

Example: What are your chances of reaching a top managerial position within the next five years?

Excellent

Good

Fair

Poor

Very poor

Tuckman draws our attention to the fact that, unlike an unstructured response which has to be coded to be useful as data, a scaled response is collected in the form of usable and analysable data.

A ‘ranking response’ is one in which a respondent is required to rank order a series of words, phrases or statements according to a particular criterion.

Example: Rank order the following people in terms of their usefulness to you as sources of advice and guidance on problems you have encountered in the classroom. Use numbers 1 to 5, with 1 representing the person most useful.

Education tutor

Subject tutor

Class teacher

Head teacher

Other student

Ranked data can be analysed by adding up the rank of each response across the respondents, thus resulting in an overall rank order of alternatives.

A ‘checklist response’ requires that the respondent selects one of the alternatives presented to her. In that they do not represent points on a continuum, they are nominal categories.

Example: I get most satisfaction in college from:

the social life

studying on my own

attending lectures

college societies

giving a paper at a seminar

This kind of response tends to yield less information than the other kinds considered.

Finally, the ‘categorical response’ mode is similar to the checklist but simpler in that it offers respondents only two possibilities.

Example: Material progress results in greater happiness for people

True      False

or:

In the event of another war, would you be prepared to fight for your country?

Yes      No

Summing the numbers of respondents with the same responses yields a nominal measure.

As a general rule, the kind of information sought and the means of its acquisition will determine the choice of response mode. Data analysis, then, ought properly to be considered alongside the choice of response mode so that the interviewer can be confident that the data will serve her purposes and analysis of them can be duly prepared. Table 21.3 summarizes the relationship between response mode and type of data.

Once the variables to be measured or studied have been identified, questions can be constructed so as to reflect them. If, for example, one of the variables was to be a new social education project that had recently been attempted with 15-year-olds in a comprehensive school, one obvious question would be: ‘How do you think the project has affected the pupils?’ Or, less directly, ‘Do you think the children have been given too much or too little responsibility?’ It is important to bear in mind that more than one question format and more than one response mode may be employed when building up a schedule. The final mixture will depend on the kinds of factors mentioned earlier – the objectives of the research, and so on.

Where an interview schedule is to be used as part of a field survey in which a number of trained interviewers are to be used, it will of course be necessary to include in it appropriate instructions for both interviewer and interviewees.

The framing of questions for a semi-structured interview will also need to consider prompts and probes (Morrison, 1993: 66). Prompts enable the interviewer to clarify topics or questions, particularly if the interviewee seems not to have understood, or to have misunderstood, or wishes to ask for clarification or more guidance from the interviewer. Probes enable the interviewer to ask respondents to extend, elaborate, add to, provide detail for, clarify or qualify their response, thereby addressing richness, depth of response, comprehensiveness and honesty that are some of the hallmarks of successful interviewing (see also Patton, 1980: 238).

A probe may be simply the follow-up ‘why’ question. It could comprise simply repeating the question, repeating the answer in a questioning tone, showing interest and understanding, asking for clarification or an example or further explication, or, indeed, simply pausing. Aldridge and Levine (2001: 119) suggest two types of probe: one in which more detailed factual information is sought, and another in which the respondent is encouraged to elaborate on accounts that they have given or opinions that they hold. Probes can range from the less intrusive (e.g. pausing for the respondent to say more, or making a sound such as ‘mmm’ to indicate that the interviewer is following closely) to the more intrusive (e.g. repeating a phrase or idea that the respondent said and then following it up with a request for further information, or summarizing (‘am I right in thinking that you were saying...’, or ‘can I just check that I have understood correctly?’) and then questioning, or asking for an example or instance, or asking for clarification, or even politely and respectfully challenging, or checking) (cf. Aldridge and Levine, 2001: 120).

I TABLE 21.3 THE SELECTION OF RESPONSE MODE

Response mode	Type of data	Chief advantages	Chief disadvantages
Fill-in	Nominal	Less biasing; greater response flexibility	More difficult to score
Scaled	Interval	Easy to score	Time-consuming; can be biasing
Ranking	Ordinal	Easy to score; forces discrimination	Difficult to complete
Checklist or Categorical	Nominal (may be interval when totalled)	Easy to score; easy to respond	Provides less data and fewer options

Source: Tuckman, 1972

Fowler (2009: 139) offers a cautionary note, suggesting that the more the interviewer prompts and probes, the greater is the chance of bias entering the interview. His argument favours standardized wording, with the possibility of further explanation if respondents are unclear.

Hence an interview schedule for a semi-structured interview (i.e. where topics and open-ended questions are written but the exact sequence and wording does not have to be followed with each respondent) might include:

   the topic to be discussed;

   the specific possible questions to be put for each topic;

   the issues within each topic to be discussed, together with possible questions for each issue;

   a series of prompts and probes for each topic, issue and question.

It would be incomplete to end this section without some comment on sampling in addition to question type, for the design of the interview has to consider who is being interviewed. ‘How many interviews do I need to conduct?’ is a frequent question of novice researchers, asking both about the numbers of people and the number of interviews with each person. The advice here echoes that of Kvale (1996: 101) that one conducts interviews with as many people as necessary in order to gain the information sought. There is no simple rule of thumb, as this depends on the purpose of the interview, for example, whether it is to make generalizations, to provide in-depth, individual data, to gain a range of responses. Though the reader is directed to Chapter 8 on sampling for fuller treatment of these matters, the issue here is that the interviewer must ensure that the interviewees selected will be able to furnish the researcher with the information, i.e. that they possess the information.

Interviewing

Setting up and conducting the interview will make up the next stage in the procedure. Where the interviewer is initiating the research herself, she will clearly select her own respondents; where she is engaged by another agent, then she will probably be given a list of people to contact. Tuckman (1972) has succinctly reviewed the procedures to adopt at the interview itself. He writes that the interviewer should inform the participant of the nature or purpose of the interview, being honest yet without risking biasing responses, and should strive to put the participant at ease. The conduct of the interview should be explained (what happens, and how, and the structure and organization of the interview), how responses may be recorded (and to seek permission if this is to happen), and these procedures should be observed throughout (cf. Fowler, 2009: 140). During the interview the biases and values of the interviewer should not be revealed, and the interviewer should avoid being judgemental. The interviewer may have to steer respondents if they are rambling off the point, without being impolite. Aldridge and Levine (2001: 119) suggest that factual, personal data should be kept until later in the interview, or at the end, rather than at the beginning of the interview.

It is crucial to keep uppermost in one’s mind the fact that the interview is a social, interpersonal encounter, not merely a data collection exercise. In this respect one has to bear in mind that different socio-cultural contexts exert different influences on an interview (not least the linguistic factor in which the researcher may be conducting the interview in a language that is not his/her first language or the respondent’s first language). Miltiades (2008) notes that, in some cultures, the influence of culture manifests itself in having not only the presence of other members of an extended family at the interview (p. 281), but in those other members actively participating in the interview, giving answers, censoring information (p. 282), interrupting, preventing information from being spoken, and passing comments, i.e. adopting a gate-keeping role (p. 283). As she remarks (p. 282), in some cultures the self is a ‘we-self’ (rather than an ‘I-self’) in a collective family, and, indeed, she notes that the Bengali language has no word for ‘private’. Just as the researcher brings his or her own cultural background to the interview, so do the respondents (p. 278), and this might affect the nature, substance and amount of data given, the possible biases towards social desirability of answers (the tendency of respondents to give what they believe will be a socially desirable response, or indeed to self-censor (p. 283)), and, indeed, in some cultures, the tendency for elders –as authority figures – to give answers rather than the initially targeted interviewees (p. 278). As she remarks (p. 281), in some cultures, the interview becomes a social event.

As the interview is a social encounter, Kvale (1996: 125) suggests that an interview follows an unwritten script for interactions, the rules for which only surface when they are transgressed. Hence the interviewer must be at pains to conduct the interview carefully and sensitively. Kvale (1996: 147) adds that, as the researcher is the research instrument, the effective interviewer is not only knowledgeable about the subject matter but is also an expert in interaction and communication. The interviewer will need to establish an appropriate atmosphere such that the participant can feel secure to talk freely. This operates at several levels.

For example there is the need to address the cognitive aspect of the interview, ensuring that the interviewer is sufficiently knowledgeable about the subject matter that she or he can conduct the interview in an informed manner, and that the interviewee does not feel threatened by lack of knowledge. That this is a particular problem when interviewing children has been documented by Simons (1982) and Lewis (1992), who indicate that children will tend to say anything rather than nothing at all, thereby limiting the possible reliability of the data. The interviewer must also be vigilant to the fact that respondents may not always be what they seem; they may be providing misinformation, telling lies, evading the issue, putting on a front (Walford, 2001: 91), settling scores and being malicious.

Further, the ethical dimension of the interview needs to be borne in mind, ensuring, for example, informed consent, guarantees of confidentiality, beneficence and non-maleficence (i.e. that the interview may be to the advantage of the respondent and will not harm her). The issues of ethics also needs to take account of what is to count as data, for example, it is often after the cassette recorder or video camera has been switched off that the ‘gems’ of the interview are revealed, or people may wish to say something ‘off the record’; the status of this kind of information needs to be clarified before the interview commences. The ethical aspects of interviewing are discussed later in the chapter.

Then there is a need to address the interpersonal, interactional, communicative and emotional aspects of the interview. For example, the interviewer and interviewee communicate non-verbally, by facial and bodily expression. Something as slight as a shift in position in a chair might convey whether the researcher is interested, angry, bored, agreeing, disagreeing and so on. Here the interviewer has to be adept at ‘active listening’.

Further, the onus is on the interviewer to establish and maintain a good rapport with the interviewee. This concerns being clear, polite, non-threatening, friendly and personable, to the point without being too assertive. It also involves being respectful, e.g. some respondents may or may not wish to be called by their first name, family name or title; being dressed too casually may not inspire confidence. Rapport also requires the interviewer to communicate very clearly and positively the purpose, likely duration, nature and conduct and contents of the interview, to give the respondent the opportunity to ask questions, to be sensitive to any emotions in the respondent, to avoid giving any signs of annoyance, criticism or impatience, and to leave the respondent feeling better than, or at least no worse than, she or he felt at the start of the interview. This requires the interviewer to put himself/herself in the shoes of the respondent, and to be sensitive to how it must feel to be interviewed. Rapport does not mean ‘liking’ the respondent (Dyer, 1995: 62); it means handling the situation sensitively and professionally.

The interviewer is also responsible for considering the dynamics of the situation, for example, how to keep the conversation going, how to motivate participants to discuss their thoughts, feelings and experiences, how to overcome the problems of the likely asymmetries of power in the interview (where the interviewer typically defines the situation, the topic, the conduct, the introduction, the course of the interview and the closing of the interview) (Kvale, 1996: 126). As Kvale suggests, the interview is not usually a reciprocal interaction between two equal participants. That said, it is important to keep the interview moving forward, and how to achieve this needs to be anticipated by the interviewer, for example by being clear on what one wishes to find out, asking those questions that will elicit the kinds of data sought, giving appropriate verbal and non-verbal feedback to the respondent during the interview. It extends even to considering when the interviewer should keep silent (p. 135).

The ‘directiveness’ of the interviewer has been scaled by Whyte (1982), where a six-point scale of directiveness and responding was devised (1 =the least directive, and 6=the most directive):

1   Making encouraging noises.

2   Reflecting on remarks made by the informant.

3   Probing on the last remark made by the informant.

4   Probing an idea preceding the last remark by the informant.

5   Probing an idea expressed earlier in the interview.

6   Introducing a new topic.

This is not to say that the interviewer should avoid being too directive or not directive enough; indeed on occasions a confrontational style might yield much more useful data than a non-confrontational style. Further, it may be in the interests of the research if the interview is sometimes quite tightly controlled, as this might facilitate the subsequent analysis of the data. For example, if the subsequent analysis will seek to categorize and classify the responses, then it might be useful for the interviewer to clarify meaning and even suggest classifications during the interview (see Kvale, 1996: 130).

Patton (1980: 210) suggests that it is important to maintain the interviewee’s motivation, hence the interviewer must keep boredom at bay, for example by keeping to a minimum demographic and background questions. The issue of the interpersonal and interactional elements reaches further, for the language of all speakers has to be considered, for example, translating the academic language of the researcher into the everyday, more easy-going and colloquial language of the interviewee, in order to generate rich descriptions and authentic data. Patton (1980: 225) goes on to underline the importance of clarity in questioning, and suggests that this entails the interviewer finding out what terms the interviewees use about the matter in hand, what terms they use amongst themselves, and avoiding the use of academic jargon. The issue here is not only that the language of the interviewer must be understandable to interviewees but that it must be part of their frame of reference, such that they feel comfortable with it.

This can be pursued even further, suggesting that the age, gender, race, class, dress, language of the interviewers and interviewees will all exert an influence on the interview itself. Bailey (1994: 183) reports that many interviewers are female, middle-class, white-collar workers, yet those they interview may have none of these characteristics. He reports that having women interviewers elicited a greater percentage of honest responses than having male interviewers (p. 182), that having white interviewers interviewing black respondents yielded different results from having black interviewers interview black respondents (pp. 180–1). He also suggests that having interviewers avoiding specific identity with particular groups or counter-cultures in their dress (e.g. rings, pins, etc.) should be eschewed (p. 185) as this can bias the interview; rather some unobtrusive clothing should be worn so as to legitimize the role of the interviewer by fitting in with the respondents’ expectations of an interviewer’s appearance. One can add here that people in power may expect to be interviewed by interviewers in powerful positions and it is more likely that an interview with a powerful person may be granted to a higher status interviewer. This is discussed fully in Chapter 9.

The sequence and framing of the interview questions will also need to be considered, for example, ensuring that easier and less threatening, non-controversial questions are addressed earlier in the interview in order to put respondents at their ease (see Patton, 1980: 210–11). This might mean that the ‘what’ questions precede the more searching and difficult ‘how’ and ‘why’ questions (though, as Patton reminds us (1980: 211), knowledge questions – ‘what’-type questions – can be threatening). The interviewer’s questions should be straightforward and brief, even though the responses need not be (Kvale, 1996: 132). He/she will also need to consider the kinds of questions to be put to interviewees, discussed earlier.

There are several problems in the actual conduct of an interview that can be anticipated and, possibly, prevented, ensuring that the interview proceeds comfortably, for example (see Field and Morse, 1989):

   avoiding interruptions from outside (e.g. telephone calls, people knocking on the door);

   minimizing distractions;

   minimizing the risk of ‘stage fright’ in interviewees and interviewers;

   avoiding asking embarrassing or awkward questions;

   jumping from one topic to another;

   giving advice or opinions (rather than active listening);

   summarizing too early or closing off an interview too soon;

   being too superficial;

   handling sensitive matters (e.g. legal matters, personal matters, emotional matters).

Arksey and Knight (1999: 53) suggest that the interviewer should:

   appear to be interested;

   keep to the interview schedule in a structured interview;

   avoid giving signs of approval or disapproval of responses received;

   be prepared to repeat questions at the respondent’s request;

   be prepared to move on to another question without irritation, if the respondent indicates unwillingness or inability to answer the question;

   ensure that he/she (the interviewer) understands a response, checking if necessary (e.g. ‘am I right in thinking that you mean...?’);

   if a response is inadequate, but the interviewer feels that the respondent may have more to say, thank the respondent and add ‘and could you please tell me...?’;

   give the respondent time to answer (i.e. avoid answering the question for the respondent).

Gadd (2004: 397) reports the significance of how the interviewer responds to the interviewee, as an unsupportive, unsympathetic or negative response (even if not intended) could discourage a respondent from proceeding.

There is also the issue of how to record the interview as it proceeds. For example, an audiotape recorder might be unobtrusive but might constrain the respondent; a videotape might yield more accurate data but might be even more constraining, with its connotation of surveillance. Merton et al. (1956) comment on the tendency of taping to ‘cool things down’. It might be less threatening not to have any mechanical means of recording the interview, in which case the reliability of the data might rely on the memory of the interviewer (though, as Gadd (2004: 384) remarks, memory is motivated in nature, and may be subject to selective recall). An alternative might be to have the interviewer make notes during the interview, but this could be highly off-putting for some respondents. The issue here is that there is a trade-off between the need to catch as much data as possible and yet to avoid having so threatening an environment that it impedes the potential of the interview situation.

What is being suggested here is that the interview, as a social encounter, has to take account of, and plan for, the whole range of other possibly non-cognitive factors that form part of everyday conduct. The ‘ideal’ interview, then, meets several ‘quality criteria’ (Kvale, 1996: 145):

   The extent of spontaneous, rich, specific and relevant answers from the interviewee.

   The shorter the interviewer’s questions and the longer the subject’s answers, the better.

   The degree to which the interviewer follows up and clarifies the meanings of the relevant aspects of the answers.

   The ideal interview is to a large extent interpreted throughout the interview.

   The interviewer attempts to verify his or her interpretations of the subject’s answers in the course of the interview.

   The interview is ‘self-communicating’ – it is a story contained in itself that hardly requires much extra descriptions and explanations.

People may refuse to be interviewed (Bailey, 1994: 186–7; Cooper and Schindler, 2001: 301), e.g. they may:

   not give a reason for refusing;

   be hostile to what they see as intrusion.

   hold anti-authoritarian feelings;

   feel that surveys are a waste of time;

   speak a foreign language;

   take an instant dislike to the interviewer;

   say that they are too busy;

   feel embarrassed or ignorant;

   dislike the topic under review;

   be afraid of the consequences of participating;

   feel inadequate or that they do not know the right answer.

The onus is on the interviewer to try to overcome these factors, whilst recognizing, of course, that they may be legitimate, in which case no further attempt can be made to conduct the interview. It is important for the interviewer to render the interview a positive, pleasant and beneficial experience, and to convince the participant of their own worth and the importance of the topic. If there is a significant difference between the interviewer and the respondent (e.g. gender, age, ethnicity, race, social status, class), then it might be advisable to have another interviewer try to conduct the interview.

So far the assumption has been that there is only one interviewer present at the interview. There is an argument for having more than one interviewer present, not only so that one can transcribe or observe features that might be overlooked by the other interviewer whilst engaging the respondent (and these roles have to be signalled clearly to the respondent at the interview), but also to share the interviewing. Joint interviews can provide two versions of the interview – a cross-check – and one can complement the other with additional points, leading to a more complete and reliable record. It also enables one interviewer to observe non-verbal features such as the power and status differentials and social dynamics, and, if there is more than one respondent present at the interview, the relationships between the respondents, e.g. how they support, influence, complement, agree and disagree with each other, or, indeed, contradict each other, the power plays at work, and so on.

On the other hand having more than one interviewer present is not without its difficulties. For example, the roles of the two interviewers may be unclear to the respondents (and it is the job of the interviewers to make this clear), or it may be intimidating to have more than one interviewer present. Researchers will need to weigh carefully the strengths and weaknesses of having more than one interviewer present, and what their roles will be.

We give readers a list of guidelines for conduct during the interview in Box 21.2.

BOX 21.2 GUIDELINES FOR THE CONDUCT OF INTERVIEWS

   Interviews are an interpersonal matter.

   Avoid saying ‘I want to know... ‘; the interviewee is doing you a favour, not being interrogated.

   How to follow up on questions/answers.

   How to keep people on track and how to keep the interview moving forward.

   How to show respect.

   How to divide your attention as interviewer and to share out the interviewees’ responses – giving them all a chance to speak in a group interview.

   Do you ask everyone in a group interview to give a response to a question?

   If there is more than one interviewer, what are the roles of the ‘silent’ interviewer, and do the interviewees know the roles of the interviewers?

   Who is looking at whom.

   If you need to look at your watch then maybe comment on this publicly.

   Try not to refer to your interview schedule; if you need to refer to it then comment on this publicly (e.g. ‘let me just check that I have covered the points that I wanted’).

   Avoid using your pen as a threatening weapon, pointing it at the interviewee.

   Consider your non-verbal communication, eye contact, signs of anxiety, showing respect.

   Give people time to think – don’t interrupt yourself if there is silence.

   How to pass over from one interviewer to another and from one interviewee to another if there is more than one interviewer or interviewee.

   How to give feedback and acceptance to the interviewees.

   Should you write responses down – what messages does this give?

   Put yourself in the shoes of the interviewee.

   What are the effects of losing eye contact or of maintaining it for too long?

   Think of your body posture – not too laid back and not too menacing.

   How to interpret and handle silence.

   Avoid looking away from the respondent if possible.

   Avoid interrupting the respondent.

   Avoid judging the respondent or his/her response.

   The interviewer should summarize and crystallize issues and build on them – that is a way of showing respect.

   How to give signs of acceptance of what people are saying, and how to avoid being judgemental.

   Take care of timing – not too long to be boring.

   Give interviewees the final chance to add any comments, and thank them at the end.

   Plan how to hand over the questions to the next interviewer.

   How to arrange the chairs and tables – do you have tables (they may be a barrier or a protection)?

   Identify who controls the data, and when the control of the data passes from the interviewee to the interviewer.

   What to do with ‘off the record’ data.

   Take time to ‘manage’ the interview and keep interviewees aware of what is happening and where it is going.

   Vary the volume/tone of your voice.

   Avoid giving your own view or opinion; be neutral.

   Who is working harder – the interviewer or the interviewee?

   Who is saying more – the interviewer or the interviewee?

   If there is more than one interviewer, how to avoid one interviewer undermining another.

   Think of prompts and probes.

   How to respond to people who say little?

   Consider the social (and physical) distance between the interviewer and interviewee(s).

   Consider the layout of the furniture – circle/oval/straight line or what?

   Have a clear introduction which makes it clear how the interview will be conducted and how the interviewees can respond (e.g. turn taking). continued

   Make sure you summarize and crystallize every so often.

   How to handle interviewees who know more about the topic than you do.

   Do you have males interviewing females and vice versa (think of age/gender/race, etc. of interviewers and interviewees)?

   Give some feedback to respondents every so often.

   What is the interview doing that cannot be done in a questionnaire?

   If there are status differentials then don’t try to alter them in the space of an interview.

   Plan what to do if the interviewee ‘turns the tables’ and tries to be the interviewer.

   Plan what to do with aggressive or angry interviewees.

   Plan what to do if powerful interviewees don’t answer your questions (maybe you need to admit that you haven’t understood very well, and ask for clarification, i.e. that it is your fault).

   Be very prepared, so that you don’t need to look at your schedule.

   Know your subject matter well.

   If people speak fast then try to slow down everything.

   As an interviewer, you have the responsibility for making sure the interview runs well.

Interviewers have to be sensitive to their own effect on the interview. For example (Cooper and Schindler, 2001: 307), they may fail to secure full cooperation or keep to procedures, they may establish an inappropriate environment (physical, cognitive, interpersonal), they may be exerting undue influence or pressure on the respondent, or they may be selective in recording the data; we consider the issue of reliability in Chapter 10.

It is important for the interviewer to explain to the respondent the purpose, scope, nature and conduct of the interview, the use to be made of the data, ethical issues, the likely duration of the interview, i.e. to explain fully the ‘rules of the game’ so that the interviewee is left in no doubt as to what will happen during and after the interview. It is important for the interviewer to introduce herself/himself properly and fully to the respondent (maybe even providing identification). The interviewer has to set the scene appropriately, for example, to say that there are no right and wrong answers, that some of the topics may be deep but that they are not designed to be a test, to invite questions and interruptions, and to clear permission for recording. During the interview it is important, also, for the interviewee to speak more than the interviewer, for the interviewer to listen attentively and to be seen by the respondent to be listening attentively, and for the interviewer to be seen to be enjoying, or at ease with, the interview.

Transcribing

This is a crucial step in interviewing, for there is the potential for massive data loss, distortion and the reduction of complexity. It has been suggested throughout that the interview is a social encounter, not merely a data collection exercise; the problem with much transcription is that it becomes solely a record of data rather than a record of a social encounter. Indeed this problem might have begun at the data collection stage, for example audiotape is selective, it filters out important contextual factors, neglecting the visual and non-verbal aspects of the interview (Mishler, 1986). Moreover, it is frequently the nonverbal communication that gives more information than the verbal communication. Morrison (1993: 63) recounts the incident of an autocratic head teacher extolling the virtues of collegiality and democratic decision making whilst shaking her head vigorously from side to side and pressing the flat of her hand in a downwards motion away from herself as if to silence discussion! To replace audio recording with video recording might make for richer data and catch nonverbal communication, but this then becomes very time-consuming to analyse.

Transcriptions inevitably lose data from the original encounter. This problem is compounded, for a transcription represents the translation from one set of rule systems (oral and interpersonal) to another very remote rule system (written language). As Kvale (1996: 166) suggests the prefix trans indicates a change of state or form; transcription is selective transformation. Therefore it is unrealistic to pretend that the data on transcripts are anything but already interpreted data. As Kvale (1996: 167) remarks, the transcript can become an opaque screen between the researcher and the original live interview situation.

Hence there can be no single ‘correct’ transcription; rather the issue becomes whether, to what extent and how a transcription is useful for the research. Transcriptions are decontextualized, abstracted from time and space, from the dynamics of the situation, from the live form, and from the social, interactive, dynamic and fluid dimensions of their source; they are frozen.

The words in transcripts are not necessarily as solid as they were in the social setting of the interview. Scheurich (1995: 240) suggests that even conventional procedures for achieving reliability are inadequate here, for holding constant the questions, the interviewer, the interviewee, the time and place does not guarantee stable, unambiguous data. Indeed Mishler (1991: 260) suggests that data and the relationship between meaning and language are contextually situated; they are unstable, changing and capable of endless reinterpretation.

We are not arguing against transcriptions, rather we are cautioning against the researcher believing that they tell everything that took place in the interview. This might require the researcher to ensure that different kinds of data are recorded in the transcript of the audiotape, for example:

   what was being said;

   the tone of voice of the speaker(s) (e.g. harsh, kindly, encouraging);

   the inflection of the voice (e.g. rising or falling, a question or a statement, a cadence or a pause, a summarizing or exploratory tone, opening or closing a line of enquiry);

   emphases placed by the speaker;

   pauses (short to long) and silences (short to long);

   interruptions;

   the mood of the speaker(s) (e.g. excited, angry, resigned, bored, enthusiastic, committed, happy, grudging);

   the speed of the talk (fast to slow, hurried or unhurried, hesitant to confident);

   how many people were speaking simultaneously;

   whether a speaker was speaking continuously or in short phrases;

   who is speaking to whom;

   indecipherable speech;

   any other events that were taking place at the same time that the researcher can recall.

If the transcript is of videotape, then this enables the researcher to comment on all of the non-verbal communication that was taking place in addition to the features noted from the audiotape. The issue here is that it is often inadequate to transcribe only spoken words; other data are important. Of course, as soon as other data are noted, this becomes a matter of interpretation (what is a long pause, what is a short pause, was the respondent happy or was it just a ‘front’, what gave rise to such-and-such a question or response, why did the speaker suddenly burst into tears). As Kvale (1996: 183) notes, interviewees’ statements are not simply collected by the interviewer, they are, in reality, co-authored.

Analysing

Once data from the interview have been collected, the next stage involves analysing them, often by some form of coding or scoring. In qualitative data the data analysis here is almost inevitably interpretive, hence the data analysis is less a completely accurate representation (as in the numerical, positivist tradition) but more of a reflexive, reactive interaction between the researcher and the decontextualized data that are already interpretations of a social encounter. The researcher has to consider whether to focus on those items that the participant mentions or reiterates the most, or whether to deem important those items that arise when the participant wanders from the point or changes the subject, or – in the case of two respondents – whether they actually mean the same even if they are using the same words to describe similar experiences or the same experience (Gadd, 2004: 385). At issue here is the unavoidable integration of analysis and interpretation.

The great tension in data analysis is between maintaining a sense of the holism of the interview and the tendency for analysis to atomize and fragment the data – to separate them into constituent elements, thereby losing the synergy of the whole, and in interviews often the whole is greater than the sum of the parts. There are several stages in analysis, for example:

   generating natural units of meaning;

   classifying, categorizing and ordering these units of meaning;

   structuring narratives to describe the interview contents;

   interpreting the interview data.

These are comparatively generalized stages. Miles and Huberman (1994) suggest 12 tactics for generating meaning from transcribed and interview data:

   counting frequencies of occurrence (of ideas, themes, pieces of data, words);

   noting patterns and themes (Gestalts), which may stem from repeated themes and causes or explanations or constructs;

   seeing plausibility – trying to make good sense of data, using informed intuition to reach a conclusion;

   clustering – setting items into categories, types, behaviours and classifications;

   making metaphors – using figurative and connotative language rather than literal and denotative language, bringing data to life, thereby reducing data, making patterns, decentring the data, and connecting data with theory;

   splitting variables to elaborate, differentiate and ‘unpack’ ideas, i.e. to move away from the drive towards integration and the blurring of data;

   subsuming particulars into the general (akin to Glaser’s (1978) notion of ‘constant comparison’ – see Chapter 33 in this book) – a move towards clarifying key concepts;

   factoring – bringing a large number of variables under a smaller number of (frequently) unobserved hypothetical variables;

   identifying and noting relations between variables;

   finding intervening variables – looking for other variables that appear to be ‘getting in the way’ of accounting for what one would expect to be strong relationships between variables;

   building a logical chain of evidence – noting causality and making inferences;

   making conceptual/theoretical coherence – moving from metaphors to constructs to theories to explain the phenomena.

This progression, though perhaps positivist in its tone, is a useful way of moving from the specific to the general in data analysis. Running through the suggestions from Miles and Huberman (1994) is the importance that they attach to coding of responses in interviews, partially as a way of reducing what is typically data overload from qualitative data.

Coding has been defined by Kerlinger (1970) as the translation of question responses and respondent information to specific categories for the purpose of analysis. As we have seen, many questions are pre-coded, that is, each response can be immediately and directly converted into a score in an objective way. Rating scales and checklists are examples of precoded questions. Coding is the ascription of a category label to a piece of data, with the category label either decided in advance or in response to the data that have been collected.

We discuss coding more fully in Chapter 33, and we refer the reader to that discussion.

Content analysis involves reading and judgement; Brenner et al. (1985) set out several steps in undertaking a content analysis of open-ended data:

1   Briefing (understanding the problem and its context in detail).

2   Sampling (of people, including the types of sample sought, see Chapter 8).

3   Associating (with other work that has been done).

4   Hypothesis development.

5   Hypothesis testing.

6   Immersion (in the data collected, to pick up all the clues).

7   Categorizing (in which the categories and their labels must: (a) reflect the purpose of the research; (b) be exhaustive; (c) be mutually exclusive).

8   Incubation (e.g. reflecting on data and developing interpretations and meanings).

9   Synthesis (involving a review of the rationale for coding and an identification of the emerging patterns and themes).

10  Culling (condensing, excising and even reinterpreting the data so that they can be written up intelligibly).

11  Interpretation (making meaning of the data).

12  Writing (including (pp. 140–3): giving clear guidance on the incidence of occurrence; proving an indication of direction and intentionality of feelings; being aware of what is not said as well as what it said – silences; indicating salience (to the readers and respondents).

13  Rethinking.

This process, the authors suggest (Brenner et al, 1985: 144), requires researchers to address several factors:

   Understand the research brief thoroughly.

   Evaluate the relevance of the sample for the research project.

   Associate their own experiences with the problem, looking for clues from the past.

   Develop testable hypotheses as the basis for the content analysis (the authors name this the ‘Concept Book’).

   Test the hypotheses throughout the interviewing and analysis process.

   Stay immersed in the data throughout the study.

   Categorize the data in the Concept Book, creating labels and codes.

   Incubate the data before writing up.

   Synthesize the data in the Concept Book, looking for key concepts.

   Cull the data, being selective is important because it is impossible to report everything that happened.

   Interpret the data, identifying its meaning and implication.

   Write up the report.

   Rethink and rewrite: have the research objectives been met?

Hycner (1985) sets out procedures that can be followed when phenomenologically analysing interview data. We saw in Chapter 1 that the phenomenologist advocates the study of direct experience taken at face value and sees behaviour as determined by the phenomena of experience rather than by external, objective and physically described reality. Hycner points out that there is a reluctance on the part of phenomenologists to focus too much on specific steps in research methods for fear that they will become reified. The steps suggested by Hycner, however, offer a possible way of analysing data which allays such fears. As he himself explains, his guidelines ‘have arisen out of a number of years of teaching phenomenological research classes to graduate psychology students and trying to be true to the phenomenon of interview data while also providing concrete guidelines’ (Hycner, 1985). In summary, the guidelines are as follows:

1   Transcription: having the interview tape transcribed, noting not only the literal statements but also non-verbal and paralinguistic communication.

2   Bracketing and phenomenological reduction: for Hycner (1985) this means, ‘suspending (bracketing) as much as possible the researcher’s meaning and interpretations and entering into the world of the unique individual who was interviewed’. The researcher thus sets out to understand what the interviewee is saying rather than what she expects that person to say.

3   Listening to the interview for a sense of the whole: this involves listening to the entire tape several times and reading the transcription a number of times in order to provide a context for the emergence of specific units of meaning and themes later on.

4   Delineating units of general meaning: this entails a thorough scrutiny of both verbal and non-verbal gestures to elicit the participant’s meaning. Hycner (1985) says, ‘It is a crystallization and condensation of what the participant has said, still using as much as possible the literal words of the participant.’ (See Box 21.3 for Hycner’s own example. This is the second page of transcription describing an experience of wonderment and awe. On the previous page, the participant discussed the background where he and his girlfriend were up in the mountains on vacation. The scene being described is the beginning of an experience of wonder.)

5   Delineating units of meaning relevant to the research question: once the units of general meaning have been noted, they are then reduced to units of meaning relevant to the research question. In the case of Hycner’s (1985) study, the original 18 general units (see Box 21.3) are reduced to 13 units of meaning relevant to the research question (see Box 21.4).

6   Training independent judges to verify the units of relevant meaning: findings can be verified by using other researchers to carry out the above procedures. Hycner’s own experience in working with graduate students well trained in this type of research is that there are rarely significant differences in the findings.

7   Eliminating redundancies: at this stage, the researcher checks the lists of relevant meaning and eliminates those clearly redundant to others previously listed.

8   Clustering units of relevant meaning: the researcher now tries to determine if any of the units of relevant meaning naturally cluster together; whether there seems to be some common theme or essence that unites several discrete units of relevant meaning. Box 21.5 gives an example of clustering units of relevant meaning.

9   Determining themes from clusters of meaning: the researcher examines all the clusters of meaning to determine if there is one (or more) central theme(s) which expresses the essence of these clusters.

10  Writing a summary of each individual interview: it is useful at this point, Hycner suggests, to go back to the interview transcription and write up a summary of the interview incorporating the themes that have been elicited from the data.

11  Return to the participant with the summary and themes, conducting a second interview: this is a check to see whether the essence of the first interview has been accurately and fully captured.

12  Modifying themes and summary: with the new data from the second interview, the researcher looks at all the data as a whole and modifies or adds themes as necessary.

13  Identifying general and unique themes for all the interviews: the researcher now looks for the themes common to most or all of the interviews as well as the individual variations. The first step is to note if there are themes common to all or most of the interviews. The second step is to note when there are themes that are unique to a single interview or a minority of the interviews.

14  Contextualization of themes: at this point it is helpful to place these themes back within the overall contexts or horizons from which these themes emerged.

15  Composite summary: Hycner considers it useful to write up a composite summary of all the interviews which would accurately capture the essence of the phenomenon being investigated. He concludes, ‘Such a composite summary describes the “world” in general, as experienced by the participants. At the end of such a summary the researcher might want to note significant individual differences’ (Hycner, 1985).

BOX 21.3 DELINEATING UNITS OF GENERAL MEANING

1 I was looking at Mary and ²all of a sudden I knew ³I was looking at her like I never looked at anybody in my whole life – and ⁴my eyes were sort of just kind of staring at her and the reason that ⁵I realized that it was tremendous was that she said to me – what are you doing – ⁶and I just said I’m looking at you – ⁷and so we just sat there and she ⁸sort of watched me look at her – and ⁹she was getting kind of uncomfortable ¹⁰and yet also kept saying – what’s going on ¹¹but not really wanting to hear – ¹²just letting me – have enough sensitivity to let me experience it – ¹³ a lot was going on – ¹⁴I didn’t realize what – what it was – ¹⁵I was just sort of sitting there-¹⁶I couldn ‘t move – ¹⁷I didn’t want to move – ¹⁸I just want to continue looking at her.

¹ Was looking at Mary

² Suddenly he knew

³ He was looking at her like he never looked at anybody in his whole life

⁴ His eyes were just staring at her

⁵ Realized it was tremendous when she said ‘What are you doing?’

⁶ He just said, ‘I’m looking at you.’

⁷ Both just sat there

⁸ She sort of watched him look at her

⁹ She was getting kind of uncomfortable

¹⁰ She kept saying ‘What’s going on?’

¹¹ She didn’t seem to want a response

¹² She had enough sensitivity to let him experience it

¹³ A lot was going on

¹⁴ He didn’t realize what was going on

¹⁵ He continued to just sit there

¹⁶ He couldn’t move

¹⁷ Didn’t want to move

¹⁸ Just wanted to continue looking at her.

Source: Hycner, 1985

Issues arising from this procedure are discussed in some detail in the second part of Hycner’s (1985) article.

BOX 21.4 UNITS OF RELEVANT MEANING

¹ Was looking at Mary

² Suddenly he knew

³ He was looking at her like he never looked at anybody in his whole life

⁴ His eyes were just staring at her

⁵ Realized it was tremendous when she said ‘What are you doing?’

⁶ He just said, ‘I’m looking at you.’

⁷ Both just sat there

¹² She had enough sensitivity to let him experience it

¹³ A lot was going on

¹⁴ He didn’t realize what was going on

¹⁵ He continued to just sit there

¹⁶ He couldn ‘t move –

¹⁷ Didn’t want to move

¹⁸ Just wanted to continue looking at her

Source: Hycner, 1985

BOX 21.5   CLUSTERS OF RELEVANT MEANING

I The tremendousness of the looking at Mary

A   Looking at Mary in a way totally different than he had ever looked at anyone in his life.^1,3

B   His eyes were just staring.⁴

C   Realized it was tremendous when she said ‘What are you doing?’⁵

D   Was (just) looking at her.⁶

E   A lot was going on.¹³

F   Just wanted to continue looking at her.¹⁶

II Realization

A   A sudden realization² (Almost like it breaks in).

B   Realized how tremendous it was (through her question).⁵

C   A lot was going on and he didn’t realize what was going on^13,14 (rhythm of awareness).

III Continuation of what was happening

A   Both just (continued) to sit there.⁷ B   He continued to sit.¹⁵

IV Inability to move

A   Couldn’t move¹⁶ (issue of volition).

B   Didn’t want to move¹⁷ (didn’t desire to move).

V Interpersonal dimension

A   Was looking at Mary in a way he had never looked at anyone in his whole life.¹³

B   Her question elicited the realization of how tremendous it was.⁵

C   He just said ‘I’m looking at you.’⁶

D   Both just sat there.⁷

Source: Hycner, 1985

Verifying

Chapter 10 discusses at length the issues of reliability, validity and generalizability of the data from interviews, and so these issues will not be repeated here. The reader is advised to explore not only that section of Chapter 10, but, indeed, the whole chapter. Kvale (1996: 237) makes the point that validation must take place at all seven stages of the interview-based investigation, set out earlier in this chapter. For example: (a) the theoretical foundation of the research must be rigorous and there must be a logical link between such theory and the research questions; (b) all aspects of the research design must be sound and rigorous; (c) the data must be accurate, reliable and valid (with consistency and reliability checks undertaken); (d) the translation of the data from an oral to a written medium must demonstrate fidelity to the key features of the interview situation; (e) data analysis must demonstrate fidelity to the data; (f) validation procedures should be in place and used; (g) the reporting should be fair and seen to be fair by readers.

One main issue here is that there is no single canon of validity; rather the notion of fitness for purpose within an ethically defensible framework should be adopted, giving rise to different kinds of validity for different kinds of interview-based research (e.g. structured to unstructured, qualitative to quantitative, nomothetic to idiographic, generalizable to unique, descriptive to explanatory, positivist to ethnographic, pre-ordinate to responsive).

Reporting

The nature of the reporting will be decided to some extent by the nature of the interviewing. For example a standardized, structured interview may yield numerical data that may be reported succinctly in tables and graphs, whilst a qualitative, word-based, open-ended interview will yield word-based accounts that take up considerably more space.

Kvale (1996: 263-6) suggests several elements of a report: (i) an introduction that includes the main themes and contents; (ii) an outline of the methodology and methods (from designing to interviewing, transcription and analysis); (iii) the results (the data analysis, interpretation and verification); (iv) a discussion.

If the report is largely numerical then figures and tables might be appropriate; if the interview is more faithfully represented in words rather than numbers then this presents the researcher with the issue of how to present particular quotations. Here Kvale (1996: 266) suggests that direct quotations should: (a) illuminate and relate to the general text whilst maintaining a balance with the main text; (b) be contextualized and be accompanied by a commentary and interpretation; (c) be particularly clear, useful and the ‘best’ of the data (the ‘gems’!); (d) should include an indication of how they have been edited; and (e) be incorporated into the natural written style of the report.

For sample interview data, see the accompanying website.

21.6 Group interviewing

One technique within the methodology of interviewing to have grown in popularity is that of group interviewing. Watts and Ebbutt (1987), for example, have considered the advantages and disadvantages of group interviewing as a means of collecting data in educational research. The advantages the authors identify include the potential for discussions to develop, thus yielding a wide range of responses. They explain, ‘ such interviews are useful ... where a group of people have been working together for some time or common purpose, or where it is seen as important that everyone concerned is aware of what others in the group are saying’ (Watts and Ebbutt, 1987). The group interview, the paper argues, can generate a wider range of responses than in individual interviews. Bogdan and Biklen (1992: 100) add that group interviews might be useful for gaining an insight into what might be pursued in subsequent individual interviews. There are practical and organizational advantages, too. Pre-arranged groups can be used for the purpose in question by teachers with minimum disruption. Group interviews are often quicker than individual interviews and hence are time-saving. The group interview can also bring together people with varied opinions, or as representatives of different collectivities.

Arksey and Knight (1999: 76) suggest that having more than one interviewee present can provide two versions of events – a cross-check – and one can complement the other with additional points, leading to a more complete and reliable record. It is also possible to detect how the participants support, influence, complement, agree and disagree with each other, and the relationships between them. On the other hand, one respondent may dominate the interview (particularly if one respondent is male and another female (p. 76)). Further, Arksey and Knight suggest that antagonisms may be stirred up at the interview, individuals may be reticent in front of others, particularly if they are colleagues or if the matter is sensitive. They also suggest that a ‘public line’ may be offered instead of a more honest, personal response, and, indeed, that participants may collude in withholding information. Watts and Ebbutt (1987) note that group interviews are of little use in allowing personal matters to emerge, or in circumstances where the researcher has to aim a series of follow-up questions at one specific member of the group. As they explain, ‘the dynamic of a group denies access to this sort of data’ (Watts and Ebbutt, 1987). Group interviews may produce ‘group think’, discouraging individuals who hold a different view from speaking out in front of the other group members. Further, Lewis (1992) comments on the problem of coding up the responses of group interviews. For further guidance on this topic and the procedures involved, we refer the reader to Simons (1982), Watts and Ebbutt (1987), Hedges (1985), Breakwell (1990), Spencer and Flin (1990), Lewis (1992) and Arksey and Knight (1999).

Several issues have to be addressed in the conduct of a group interview, for example:

1   How to divide your attention as interviewer and to share out the interviewees’ responses – giving them all a chance to speak in a group interview?

2   Do you ask everyone in a group interview to give a response to a question?

3   How to handle people who are too quiet, too noisy, who monopolize the conversation, who argue and disagree with each other?

4   What happens if people become angry with you or with each other?

5   How to make people be quiet/stop talking whilst being polite?

6   How to handle differences in how talkative people are?

7   How to arrange turn-taking (if appropriate)?

8   Do you ask named individuals questions?

9   How can you have individuals answer without forcing them?

10   How to handle a range of very different responses to the same question?

11   Why have you brought together the particular people in the group?

12   Do you want people to answer in a particular sequence?

13   What to do if the more experienced people always answer first in a group interview?

14   As an interviewer, be vigilant to pick up on people who are trying to speak.

It must be borne in mind when conducting group interviews that the unit of analysis is the view of the whole group and not the individual member; a collective group response is being sought, even if there are individual differences or a range of responses within the group. This ensures that no individual is either unnecessarily marginalized or subject to blame or being ostracized for holding a different view.

Group interviews are also very useful when interviewing children, and it is to this that we now turn.

21.7 Interviewing children

Children have been regarded as ‘the best sources of information about themselves’ (Docherty and Sandelowski, 1999: 177), but it is important for the interviewer to be able to enter their world and childhood culture and to see the situation through their eyes (Docherty and Sandelowski, 1999: 177). It is important to understand the world of children through their own eyes rather than the lens of the adult. Children differ from adults in cognitive and linguistic development, attention and concentration span, ability to recall, life experiences, what they consider to be important, status and power (Arksey and Knight, 1999: 116). All these have a bearing on the interview. Arksey and Knight also indicate (pp. 116–18) that it is important to establish trust with children, to put the child at ease quickly and to help him/her to feel confident, to avoid overreacting (e.g. if the child is distracted), to make the interview non-threatening and enjoyable, to use straightforward language and child’s language, to ask questions that are appropriate for the age of the child (e.g. to keep to the ‘here and now’, to avoid using ‘why’, ‘when’ and ‘how’ questions with very young children, e.g. below five years old), to ensure that children can understand abstract questions (often for older children), to allow time to think, and to combine methods and activities in an interview (e.g. drawing, playing, writing, speaking, playing a game, using pictures, newspapers, toys or photographs).

Group interviewing can be useful with children, as it encourages interaction between the group rather than simply a response to an adult’s question. Group interviews of children might also be less intimidating for them than individual interviews (Greig and Taylor, 1999: 132). Eder and Fingerson (2003: 34) suggest that a power and status dynamic is heavily implicated in interviewing children; they have little in comparison to the adult. Indeed, Thorne (1994) uses the term ‘kids’ rather than ‘children’, as the former is the term used by the children themselves, whereas ‘children’, she argues, is a term used exclusively by adults, denoting subordinacy (cf. Eder and Fingerson, 2003: 34). Mayall (1999) suggests regarding children as a ‘minority group’, in that they lack power and control over their own lives. If this is the case, then it is important to take steps to ensure that children are given a voice and an interview setting in which they feel comfortable (cf. Maguire, 2005). Group interviewing is such a setting, taking place in as close to a natural surrounding as possible (Greig and Taylor, 1999: 131); indeed Eder and Fingerson (2003: 45) report the successful use of a high status child as the interviewer with a group of children.

Group interviewing with children enables them to challenge each other and participate in a way that may not happen in a one-to-one, adult-child interview and using language that the children themselves use. For example, Lewis (1992) found that ten-year-olds’ understanding of severe learning difficulties was enhanced in group interview situations, the children challenging and extending each other’s ideas and introducing new ideas into the discussion. Further, having the interview as part of a more routine, everyday activity can also help to make it less unnatural, as can making the interview more like a game (e.g. by using props such as toys and pictures). For example, it could be part of a ‘show and tell’ or ‘circle time’ session, or part of group discussion time. The issue here is to try to make the interview as informal as possible. Of course, sometimes it may be more useful to formalize the session, so that children have a sense of how important the situation is, and they can respond to this positively. It can be respectful to have an informal or, indeed, a formal interview; the former maybe for younger children and the latter for older children.

Whilst group interviews may be useful with many children, it is also the case that individual interviews with children may also be valuable. For example, Eder and Fingerson (2003: 43–4) report the value of individual interviews with adolescents, particularly about sensitive matters, for example relationships, family, body issues, sexuality, love. Indeed they report examples where individual interviews yielded different results from group interviews with the same people about the same topics, and where the individuals valued greatly the opportunity for a one-to-one conversation.

Interviews with children should try to employ open-ended questions, to avoid a single answer type of response (Greig and Taylor, 1999; Wright and Powell, 2006), as answers to open-ended questions are usually more accurate than answers to closed questions (Wright and Powell, 2006: 317) since they are respondent-driven and respondent-focused, and they can take greater account of children with limited linguistic or cognitive abilities. Closed questions can lead to response bias in that children may provide answers without thinking (Wright and Powell, 2006: 317). Waterman et al. (2001) report that children gave clear answers to yes/no closed questions even when such question types were deliberately given in respect of unanswerable questions (i.e. where insufficient information had been given for the question to be answered), in other words, the format of the question artificially skewed the response. Clearly, however, specific questions may be needed to elicit specific, e.g. factual, details (Wright and Powell, 2006: 320).

Another strategy for interviewing children is to use a projection technique. Here, instead of asking direct questions, the interviewer can show a picture or set of pictures, and then ask the children for their responses to it/them (cf. Greig and Taylor, 1999: 132–3). For example, a child may first comment on the people’s race in the pictures, followed by their sex, suggesting that race may be more important in their mind than their sex. This avoids a direct question and may reduce the possibility of a biased answer – where the respondent may be looking for cues as to how to respond. Other projection techniques include the use of dolls or puppets, photographs of a particular scene which the respondents have to comment upon (e.g. what is happening? What should be done here?) and the ‘guess who?’ technique (Wragg, 2002: 157) (which people might fit a particular description).

Simons (1982), Lewis (1992), Bailey (1994: 447–9), Breakwell (2000: 245–6) and Breakwell et al. (2006: 245–6), however, chart some difficulties in interviewing children, for example how to:

   overcome children being easily distracted (e.g. some interviewers provide toys or pictures, or children may be fascinated by something as simple as the researcher’s pen, or there may be a passing vehicle outside, and these distract the children);

   avoid the researcher being seen as an authority figure (e.g. a teacher, a parent or an adult in a powerful position);

   understand what children mean and what they say (particularly with very young children);

   gather a lot of information in a short time, children’s attention span being limited;

   have children reveal what they really think and feel rather than what they think the researcher wants to hear;

   avoid the situation being seen by the child as a test;

   keep the interview relevant;

   overcome young children’s unwillingness to contradict an adult or assert themselves, or, in the case of adolescents, deliberately being oppositional in their views;

   interview inarticulate, hesitant and nervous children;

   get the children’s teacher away from the children;

   respond to the child who says something then immediately wishes she hadn’t said it;

   elicit genuine responses from children rather than simply responses to the interview situation;

   get beyond the institutional, head teacher’s or ‘expected’ response;

   avoid receiving a socially desirable response;

   ensure that the child is giving a true opinion;

   keep children to the point;

   avoid children being too extreme or destructive of each other’s views;

   pitch language at the appropriate level;

   overcome the children (particularly young children) taking a question too literally (hence the need to avoid metaphors, similes or analogies);

   enable the children to see a situation through other people’s eyes;

   avoid the interview being an arduous bore;

   overcome children’s poor memories;

   avoid children being too focused on particular features or situations;

   avoid the situation where the child will say ‘yes’ to anything (an ‘acquiescence bias’) addressed, for example, by avoiding ‘yes/no’ questions in favour of open-ended questions;

   overcome the situation of the child saying anything in order to please;

   overcome the unwillingness of children to contradict an adult or to assert themselves in the presence of an adult;

   overcome the proclivity of some children to say that they ‘don’t know’ (for a variety of reasons, e.g. they are not interested, they genuinely don’t know, they don’t understand the question, they think that the interviewer might expect them not to know, they are unwilling to disclose what they do know, they are too shy to speak, they cannot explain themselves very well), or simply to shrug their shoulders and remain silent;

   overcome the problem that some children will say anything rather than feel they do not have ‘the answer’;

   overcome the problem that some children dominate the conversation;

   avoid the problem of children feeling very exposed in front of their friends;

   avoid children feeling uncomfortable or threatened (addressed, perhaps, by placing children with their friends);

   avoid children telling lies.

Clearly these problems are not exclusive to children; they apply equally well to some adult group interviews. Group interviews require skilful chairing and attention to the physical layout of the room so that everyone can see everyone else. Group size is also an issue; too few and it can put pressure on individuals, too large and the group fragments and loses focus. Lewis (1992) summarizes research to indicate that a group of around six or seven is an optimum size, though it can be smaller for younger children. The duration of an interview may not be for longer than, at most, 15 minutes, and it might be useful to ensure that distractions are kept to a minimum. Simple language to the point and without ambiguity (e.g. avoiding metaphors) is important. It is crucial to keep in mind that an interview is a social encounter, and children may be very sensitive to the social dynamics and social context of the interview (Morison et al, 2000), and not only its cognitive element (Maguire (2005: 4) suggests that ‘children have good social radar’. Children will be sensitive to the sex and ethnicity of the interviewer; the very fact that the interviewer is an adult will affect the interview (Morison et al, 2000: 113). For further information we refer the reader to Wilson and Powell (2001) and Mukerji and Albon (2010).

21.8 Interviewing minority and marginalized people

Not all the methods of interviewing set out so far will apply to interviewing marginalized people, i.e. those who are ‘on the edge’ of society (Barron, 1999), e. g. economically, socially, politically, or who are ‘invisible’: stigmatized groups, the unemployed, the elderly, refugees, asylum seekers, travellers, those with special needs, those with a low status in society, those with limited linguistic, cognitive or intellectual abilities, those whose first language is a minority language, the disabled, the chronically ill, children with cerebral palsy, victims of crime, the oppressed, the subordinate and so on. That this is an area of increasing educational concern is attested by Parker and Lynn (2002: 13), who argue that much educational research has ignored marginalized groups by not addressing their concerns or including them as areas of research, or that regard them as minorities that do not merit research, and, where, if research is conducted with/on them, it uses culturally inappropriate methods of investigation (p. 13). Similarly Kelly (2007: 22) argues that researchers can no longer ‘exclude learning disabled children’ from research on the grounds that they pose challenges to conventional research methods.

In interviewing marginalized groups, the interviewer will need to consider greater use of informal, open-ended interviews (which follow the train of thought and response of the respondent and which use age-appropriate and context-appropriate language) rather than highly structured interviews (Swain et al, 1998: 26). Further, the authors recommend the use of narrative, qualitative and in-depth interviews (discussed later in this chapter), enabling self-disclosure (both by the interviewer and the interviewee) (Swain et al, 1998: 26), wherein participants ‘tell their stories’ (Barron, 1999: 38) in their own words, and recount their subjective experiences and feelings. This gives them a ‘voice’, where otherwise they would either not be heard or listened to (see also Swain et al, 1998). This accords with the emancipatory potential or intent of research that was set out as a key feature of critical educational research in Chapter 2 (see also Barron, 1999: 40, 44–6) Barron suggests that it is important for the respondent to feel safe, secure, supported, close to the interviewer, and to know that he/she has the undivided attention of the interviewer (p. 41) and a non-judgemental and non-evaluative stance by the interviewer, with built-in opportunities for respondent validation and clearance (respondents may wish, upon reflection, to withdraw comments initially made at interview).

The researcher must take care not to exploit what are likely to be (perceived) asymmetries of power in the interview (where the researcher may be regarded as having more than the respondent) (Swain et al, 1998: 26). Indeed Swain et al (1998: 25) remind researchers that the interviews and research on marginalized groups should bring benefit to them, i.e. they do not continue to be exploited or marginalized.

An interview is a communicative encounter, and, for some marginalized groups (e.g. the physically disabled or those with communication difficulties) this is precisely the challenge to be faced by researchers: how to communicate with those who cannot communicate easily or at all (e.g. those who cannot speak, elective mutes, the deaf, children with degrees of autism, children with Down’s syndrome or attention deficit disorder). Here Kelly (2007: 25–6) notes the use of communication cards, pictorial cards (e.g. that indicate feelings), drawing frames, picture books, toys, puppets, photographs of familiar people or places, respecting and working with – and in – the language used by the participant and keeping within their frame of reference, considering the greater use of yes/no questions than open questions (e.g. for students who cannot speak but who can point). She also advocates the use of projection techniques (p. 28) such as ‘three wishes’, asking participants to draw a picture to represent the matter in hand, and the use of ‘feeling cards’ (cards with pictures of feelings).

Kelly (2007: 24) comments that gaining access to marginalized groups may be difficult, and that, in the case of those with disabilities, it is likely to be necessary to gain access through gatekeepers, e.g. parents, social workers, health team members, carers and the suchlike, and, indeed, to have them present during the interview or to speak on the respondents’ behalf (e.g. to protect the respondents’ rights directly or in acting as proxies/advocates). This is important, as Kelly (p. 23) reports the dangers of ‘suggestibility’ of participants in interviews, and it leads to her comment that skilful and sensitive questioning are essential, drawing on participants’ actual experiences, and taking care not to project the researcher’s own interests. Bourne-Day and Lee-Treweek (2008) also indicate that issues of privacy and identity may be highly significant in researching marginalized groups.

In conducting interviews, Kelly (2007) makes the point that it may be necessary to hold several short interviews rather than a single long interview, in order for the participants to be able to concentrate, retain their attention (p. 25) and not become tired. She also emphasizes the importance of waiting longer for an answer to be given, and to be alert to different ways in which children can communicate other than through speech (p. 28), e.g. facial expression, writing, signing, gestures and non-verbal communication, symbols (see also Mitchell and Sloper, 2008: 11), drawing and game playing. There has to be a shift, Kelly avers, away from a deficit model in which children cannot speak, and towards a positive model of how they can communicate through other means.

Morgan (1996) suggests that interviewing marginalized groups can be addressed usefully through group interviewing and with focus groups, and it is to focus groups that we turn now.

21.9 Focus groups

As an adjunct to group interviews, the use of focus groups is growing in educational research, albeit more slowly than, for instance, in business and political circles. Focus groups are a form of group interview, though not in the sense of a backwards and forwards between interviewer and group. Rather, the reliance is on the interaction within the group who discuss a topic supplied by the researcher (Morgan, 1988: 9), yielding a collective rather than an individual view. Hence the participants interact with each other rather than with the interviewer, such that the views of the participants can emerge – the participants’ rather than the researcher’s agenda can predominate. It is from the interaction of the group that the data emerge. Focus groups are contrived settings, bringing together a specifically chosen sector of the population, previously unknown to each other (Hydén and Büllow, 2003) to discuss a particular given theme or topic, where the interaction with the group leads to data and outcomes (Smithson, 2000; Hydén and Büllow, 2003). Their contrived nature is both their strength and their weakness: they are unnatural settings yet they are very focused on a particular issue and, therefore, will yield insights that might not otherwise have been available in a straightforward interview; they are economical on time, producing a large amount of data in a short period of time, but they tend to produce less data than interviews with the same number of individuals on a one-to-one basis (Hydén and Bülow, 2003: 19).

Focus groups (Morgan, 1988; Krueger, 1988; Bailey, 1994: 192–3; Robson, 2002: 284–5) are useful for:

   orientation to a particular field of focus;

   developing themes, topic and schedules flexibly for subsequent interviews and/or questionnaires;

   generating hypotheses that derive from the insights and data from the group;

   generating and evaluating data from different sub-groups of a population;

   gathering qualitative data;

   generating data quickly and at low cost;

   gathering data on attitudes, values and opinions;

   empowering participants to speak out, and in their own words;

   encouraging groups, rather than individuals, to voice opinions;

   encouraging non-literate participants;

   providing greater coverage of issues than would be possible in a survey;

   gathering feedback from previous studies.

Focus groups might be useful to triangulate with more traditional forms of interviewing, questionnaire, observation, etc. There are several issues to be addressed in running focus groups, for example (Morgan, 1988: 41–8):

   deciding the number of focus groups for a single topic (one group is insufficient, as the researcher will be unable to know whether the outcome is unique to the behaviour of the group);

   deciding the size of the group (too small and intra-group dynamics exert a disproportionate effect, too large and the group becomes unwieldy and hard to manage; it fragments). Morgan (1988: 43) suggests between four and 12 people per group, whilst Fowler (2009: 117) suggests between six and eight people;

   how to allow for people not ‘turning up’ on the day. Morgan (1988: 44) suggests the need to over-recruit by as much as 20 per cent;

   taking extreme care with the sampling, so that every participant is the bearer of the particular characteristic required or that the group has homogeneity of background in the required area, otherwise the discussion will lose focus or become unrepresentative. Sampling is a major key to the success of focus groups;

   ensuring that participants have something to say and feel comfortable enough to say it;

   chairing the meeting so that a balance is struck between being too directive and veering off the point, i.e. keeping the meeting open-ended but to the point.

Newby (2010: 350–1) indicts that focus groups should be clear on the agenda and the focus, take place in a setting that is conducive to discussion, have a skilled moderator who can prompt people to speak, promote thinking and reflection, and should have a record kept.

Unlike group interviewing with children, discussed above, focus groups operate more successfully if they are composed of relative strangers rather than friends unless friendship, of course, is an important criterion for the focus (e.g. that the group will discuss something that is usually only discussed amongst friends).

Focus groups are not without their drawbacks. For example they tend not to yield numerical, quantifiable or generalizable data; the data may be difficult to analyse succinctly; the number of people involved tends to be small; they may yield less information than a survey; and the group dynamics may lead to non-participation by some members and dominance by others (e.g. status differentials may operate); the number of topics to be covered may be limited; intra-group disagreement and even conflicts may arise; inarticulate members may be denied a voice; the data may lack overall reliability. Further, Smithson (2000) suggests that there is a problem of only one voice being heard, particularly if there is a dominant member of the group, and for the group dynamics to suppress dissenting voices or different views on controversial topics, even though the group moderator may try to prevent this.

Although its potential is considerable, the focus group, as a particular kind of group interviewing, still has to find its way into educational circles to the extent that it has in other areas of life. Focus groups require skilful facilitation and management by the researcher.

21.10 Non-directive, focused, problem-centred and in-depth interviews

Originating from psychiatric and therapeutic fields with which it is most readily associated, the non-directive interview is characterized by a situation in which the respondent is responsible for initiating and directing the course of the encounter and for the attitudes she expresses in it (in contrast to the structured or research interview we have already considered, where the dominating role assumed by the interviewer results in, to use Kitwood’s (1977) phrase, ‘an asymmetry of commitment’). It has been shown to be a particularly valuable technique because it gets at the deeper attitudes and perceptions of the person being interviewed in such a way as to leave them free from interviewer bias. We shall examine briefly the characteristics of the therapeutic interview and then consider its usefulness as a research tool in the social and educational sciences.

The non-directive interview as it is currently understood grew out of the pioneering work of Freud and subsequent modifications to his approach by later analysts. His basic discovery was that if one can arrange a special set of conditions and have a patient talk about his difficulties in a certain way, behaviour changes of many kinds can be accomplished. The technique developed was used to elicit highly personal data from patients in such a way as to increase their self-awareness and improve their skills in self-analysis. By these means they became better able to help themselves. As Madge (1965) observes, it is these techniques which have greatly influenced contemporary interviewing techniques, especially those of a more penetrating and less quantitative kind.

The present-day therapeutic interview has its most persuasive advocate in Carl Rogers who has on different occasions testified to its efficacy. Basing his analysis on his own clinical studies, he has identified a sequence of characteristic stages in the therapeutic process, beginning with the client’s decision to seek help. He is met by a counsellor who is friendly and receptive, but not didactic. The next stage is signalled when the client begins to give vent to hostile, critical and destructive feelings, which the counsellor accepts, recognizes and clarifies. Subsequently, and invariably, these antagonistic impulses are used up and give way to the first expressions of positive feeling. The counsellor likewise accepts these until suddenly and spontaneously ‘insight and self-understanding come bubbling through’ (Rogers, 1942). With insight comes the realization of possible courses of action and also the power to make decisions. It is in translating these into practical terms that the client frees himself from dependence on the counsellor.

Rogers (1945) subsequently identified a number of qualities in the interviewer which he deemed essential: that she bases her work on attitudes of acceptance and permissiveness; that she respects the client’s responsibility for his own situation; that she permits the client to explain his problem in his own way; and that she does nothing that would in any way arouse the client’s defences.

Such then are the principal characteristics of the non-directive interview technique in a therapeutic setting. But what of its usefulness as a purely research technique in societal and educational contexts? There are a number of features of the therapeutic interview which are peculiar to it and may well be inappropriate in other settings: for example, as we have seen, the interview is initiated by the respondent; his motivation is to obtain relief from a particular symptom; the interviewer is primarily a source of help, not a procurer of information; the actual interview is part of the therapeutic experience; the purpose of the interview is to change the behaviour and inner life of the person and its success is defined in these terms; and there is no restriction on the topics discussed.

A researcher has a different order of priorities, however, and what appear as advantages in a therapeutic context may be decided limitations when the technique is used for research purposes, even though she may be sympathetic to the spirit of the non-directive interview. As Madge (1965) explains, increasingly there are those ‘who wish to retain the good qualities of the non-directive technique and at the same time are keen to evolve a method that is economical and precise enough to leave a residue of results rather than merely a posse of cured souls’.

One attempt to meet this need is to be found in a programme reported by Merton and Kendall (1946) in which the focused interview was developed. While seeking to follow closely the principle of non-direction, the method did introduce rather more interviewer control in the kinds of questions used and sought also to limit the discussion to certain parts of the respondent’s experience.

The focused interview differs from other types of research interview in certain respects. These have been identified by Merton and Kendall (1946) as follows:

1   The persons interviewed are known to have been involved in a particular situation: they may, for example, have watched a TV programme; or seen a film; or read a book or article; or have been a participant in a social situation.

2   By means of the techniques of content analysis, elements in the situation which the researcher deems significant have previously been analysed by her. She has thus arrived at a set of hypotheses relating to the meaning and effects of the specified elements.

3   Using her analysis as a basis, the investigator constructs an interview guide. This identifies the major areas of enquiry and the hypotheses which determine the relevant data to be obtained in the interview.

4   The actual interview is focused on the subjective experiences of the people who have been exposed to the situation. Their responses enable the researcher both to test the validity of her hypotheses, and to ascertain unanticipated responses to the situation, thus giving rise to further hypotheses.

From this it can be seen that the distinctive feature of the focused interview is the prior analysis by the researcher of the situation in which subjects have been involved. The advantages of this procedure have been cogently explained by Merton and Kendall:

Foreknowledge of the situation obviously reduces the task confronting the investigator, since the interview need not be devoted to discovering the objective nature of the situation. Equipped in advance with a content analysis, the interviewer can readily distinguish the objective facts of the case from the subjective definitions of the situation. He thus becomes alert to the entire field of ‘selective response’. When the interviewer, through his familiarity with the objective situation, is able to recognize symbolic or functional silences, ‘distortions’, avoidances, or blockings, he is the more prepared to explore their implications.

(Merton and Kendall, 1946: 541)

In the quest for what Merton and Kendall (1946) term ‘significant data’, the interviewer must develop the ability to evaluate continuously the interview while it is in progress. To this end, they established a set of criteria by which productive and unproductive interview material can be distinguished. Briefly, these are:

1   Non-direction: interviewer guidance should be minimal.

2   Specificity: respondents’ definitions of the situation should find full and specific expression.

3   Range and scope: the interview should maximize the range of evocative stimuli and responses reported by the subject.

4   Depth and personal context: the interview should bring out the affective and value-laden implications of the subjects’ responses, to determine whether the experience has central or peripheral significance. It should elicit the relevant personal context, the idiosyncratic associations, beliefs and ideas.

Witzel (2000) advocates the use of the problem-centred interview for gathering objective evidence on human behaviour and subjective views on social phenomena. He indicates three principles underpinning the problem-centred interview:

   a ‘problem-centred orientation’ toward socially relevant problems (p. 2);

   methodological flexibility (e.g. group interviews, individual interviews, the biographical interview, structured and less structured interviews) in the ‘object-orientation’ (i.e. in order to address different kinds of problem) (p. 3);

   a ‘process orientation’, i.e. attempting to reconstruct the actions and orientations of the participant (p. 3).

The problem-centred interview, Witzel (2000) avers, can use: (a) a structured, short questionnaire at interview in order to gather factual data about the participants (e.g. age, sex, occupation, education); (b) an interview schedule (guidelines) in order to structure the interview, with lead questions; (c) recording equipment to ensure accuracy of the account and to avoid the interviewers having to take notes (i.e. able to concentrate on the discussion); and (d) a postscript (pp. 3–4), written directly after the interview, to contain reflections, key points, observations and interpretations.

The in-depth interview, as its name suggests, is conducted to explore issues, personal biographies, and what is meaningful to, or valued by, participants, how they feel about particular issues, how they look at particular issues, their attitudes, opinions and emotions (cf. Newby, 2010: 243–4). They tend to be semi-structured, to enable the course of the respondents’ responses to dictate the direction of the interview, though the researcher also has an interview schedule to keep an interview on track, and may operate probes to enquire further into issues. They may feature in case studies, action research and, as the work of Ball (1990, 1994a, 1994b) and Bowe et al. (1992) testify, they may feature in interviewing powerful people and policy makers. On the other hand, they may be useful in gathering data from marginalized or stigmatized groups in society (Newby (2010: 243–4) gives the example of migrants and refugees). Given the intensive and extensive nature of these interviews, gaining access and permission may be difficult, not least as they may take time to conduct.

21.11 Telephone interviewing

We advise the reader to take this section in conjunction with Chapter 13 on telephone interviews in surveys. The use of telephone interviewing has long been recognized as an important method of data collection and is common practice in survey research, though, as Arksey and Knight (1999: 79) aver, telephone interviews do not feel like interviews, as both parties are deprived of several channels of communication and the establishment of a positive relationship (e.g. non-verbal). We explore this here. Dicker and Gilbert (1988), Nias (1991), Oppenheim (1992), Borg and Gall (1996), Shaughnessy et al. (2003) and Shuy (2003) suggest several attractions to telephone interviewing:

   It is sometimes cheaper and quicker than face-to-face interviewing.

   It enables researchers to select respondents from a much more dispersed population than if they have to travel to meet the interviewees.

   Travel costs are omitted.

   It is particularly useful for brief surveys.

   It may protect the anonymity of respondents more than a personal interview.

   It is useful for gaining rapid responses to a structured questionnaire.

   Monitoring and quality control are undertaken more easily since interviews are undertaken and administered centrally, indeed there are greater guarantees that the researcher actually carries out the interview as required.

   Interviewer effects are reduced.

   There is greater uniformity in the conduct of the interview and the standardization of questions.

   There is greater interviewer control of the interview.

   The results tend to be quantitative.

   They are quicker to administer than face-to-face interviews because respondents will only usually be prepared to speak on the telephone for, at most, 15 minutes.

   Callback costs are so slight as to enable frequent callbacks, enhancing reliability and contact.

   Many groups, particularly of busy people, can be reached at times more convenient to them than if a visit were to be made.

   They are safer to undertake than, for example, having to visit dangerous neighbourhoods.

   They can be used to collect sensitive data, as possible feelings of threat of face-to-face questions about awkward, embarrassing or difficult matters are absent.

   It does not rely on the literacy of the respondent (as, for example, in questionnaires).

   The use of the telephone may put a little pressure on the respondent to respond, and it is usually the interviewer rather than the interviewee who terminates the call.

   Response rate is higher than, for example, questionnaires.

More recently, the use of smartphones in interviewing has increased (Raento et al, 2009), and is a powerful tool, as increasing numbers of people carry them around on a permanent basis and hence can be contacted easily, regardless of their location. Further, with increasing additional functionality, the smartphone offers a potentially powerful tool for researchers. Smartphones, the authors (Raento et al, 2009: 428) aver, have the attraction of flexibility (many uses and computer-like functions on a single smartphone, e.g. video and image recording for (self-) documentation, and cost-efficiency; they enable easy access and relatively unobtrusive data collection as data can be collected in real time or stored (increasing ecological validity); and they enable high granularity of data to be gathered (p. 442)).

Taking forward the use of electronic media, James and Busher (2007) and James (2007) argue for the power of email interviewing as a qualitative method in educational research, as it enables the researcher to contact hard-to-reach groups (e.g. by virtue of practical constraints such as time and availability of both parties to meet face-to-face, location and travelling, geographical dispersion, disability and language or communication (James, 2007: 966)) (see also Bampton and Cowton, 2002). Indeed James and Busher (2007: 405) argue that email interviews can generate fuller, richer, more reflective, thoughtful and longer answers than telephone interviewing.

Email interviews also reduce transcription time as the email is already transcribed, and the interviewee, therefore, has the opportunity to check what data are being given, thereby overcoming the possibility of respondents in a face-to-face interview saying something they later wish to withdraw (Bampton and Cowton, 2002: 4). Whether or not the absence of face-to-face contact, the visibility of the participants and the absence of non-verbal cues increases or reduces reliability is a moot point (p. 969). Similarly, the researcher will need to make continual efforts to ensure that the respondent:

   keeps to the point;

   fully understands the nature, focus and purpose of the interview;

   knows the number of questions that will be asked (particularly if there are several email exchanges);

   knows that they should not delete previous emails that are part of the interview;

   knows the time frame in which to reply to an email (cf. James and Busher, 2006: 407–9).

Email interviews can be conducted synchronously, in real time, or asynchronously (James and Busher, 2006: 970) (the latter can afford the interviewee some time to consider his/her responses (Bampton and Cowton, 2002: 3), and they ‘democratise narrative exchanges’ between the interviewer and the respondent (p. 970) (though James and Busher (2007: 416) contest this latter point). That said, Bampton and Cowton (2002: 5) caution against bombarding the respondent with too many questions in a single interview; rather, they suggest, the questions could be spaced over more than one email. Further, they indicate the need to signal to the respondent when the email interview is nearing its close. E-interviewing is susceptible to technological problems (e.g. unstable connectivity, slow connections (particularly in video-conferencing) mailbox being full), and these must be explored before the e-interview is conducted.

Telephone interviewing is not as cut-and-dried as the claims made for it, as there are several potential problems with telephone interviewing, for example:

   It is very easy for respondents simply to hang up on the caller.

   Motivation to participate may be lower than for a personal interview.

   There is a chance of skewed sampling, as not all of the population have a telephone (often those lower income households – perhaps the very people that the researcher wishes to target) or can hear (e.g. the old and second language speakers in addition to those with hearing difficulties).

   There is a lower response rate at weekends.

   The standardized format of telephone interviews may prevent thoughtful or deep answers from being provided.

   Some people have a deep dislike of telephones, that sometimes extends to a phobia, and this inhibits their responses or willingness to participate.

   Respondents may not disclose information because of uncertainty about actual (even though promised) confidentiality.

   Respondents may come to snap judgements without the adequate or deeper reflection necessary for a full answer to serious issues.

   Respondents may not wish to spend a long time on the telephone, so telephone interviews tend to be briefer than other forms of interview.

   Concentration spans are shorter than in a face-to-face interview.

   The interviewer has to remain bright and focused, listen very carefully and respond – it is tiring.

   Questions tend to be closed, fixed and simple.

   There is a limit on the complexity of the questions that can be put.

   Response categories must be very simple or else respondents will forget what they are.

   Many respondents (up to 25 per cent (Oppenheim, 1992: 97)) will be ‘ex-directory’ and so their numbers will not be available in telephone directories.

   Respondents may withhold important information or tell lies, as the non-verbal behaviour that frequently accompanies this is not witnessed by the interviewer.

   It is often more difficult for complete strangers to communicate by telephone than face-to-face, particularly as non-verbal cues are absent.

   Respondents are naturally suspicious (e.g. of the caller trying to sell a product).

   One telephone might be shared by several people.

   Some respondents feel that telephone interviews afford less opportunity for them to question or rebut the points made by the interviewer.

   There may be distractions for the respondent (e.g. a television may be switched on, children may be crying, others may be present).

   Responses are difficult to write down or record during the interview.

That said, Sykes and Hoinville (1985) and also Borg and Gall (1996) suggest that telephone interviewing reaches nearly the same proportion of many target populations as ‘standard’ interviews, that it obtains nearly the same rate of response, and produces comparable information to ‘standard’ interviews, sometimes at a fraction of the cost. The response rate issue is contested: Weisberg et al. (1996: 122) and Shuy (2003: 181) report lower response rates to telephone interviews.

Harvey (1988), Oppenheim (1992) and Miller (1995) consider that: (a) telephone interviews need careful arrangements for timing and duration (typically that they are shorter and quicker than face-to-face interviews) – a preliminary call may be necessary to fix a time for a longer call to be made; (b) the interviewer will need to have ready careful prompts and probes, including more than usual closed questions and less complex questions, in case the respondent ‘dries up’ on the telephone; (c) both interviewer and interviewee need to be prepared in advance of the interview if its potential is to be realized; and that (d) sampling requires careful consideration, using, for example, random numbers or some form of stratified sample. In general, however, many of the issues from ‘standard’ forms of interviewing apply equally well to telephone interviewing. Further, Houtkoop-Steenstra and van den Bergh (2000) report that an agenda-based introduction (in which interviewers formulated their own introductions based on a small number of key words) is more effective in securing higher response rates than standardized, scripted introductions.

Face-to-face interviews may be more suitable than telephone interviews (Weisberg et al., 1996: 122; Shuy, 2003: 179–82) if: (a) the interviewer wishes to address complex issues or sensitive questions; (b) a natural context might yield greater accuracy; (c) deeper and self-generated answers are sought (i.e. where the question does not frame the answer too strongly); (d) issues requiring probing, deep reflection and, thereby, a longer time is sought; (e) greater equality of power between interviewer and respondent is sought; (f) older, second language speakers and hearing-impaired respondents are being interviewed; (g) marginalized respondents are being sought.

It is not uncommon for telephone interviewing to be outsourced, and this might be an advantage or a disadvantage. On the one hand it takes pressure off the researcher, not only because of the time involved but also because a 15-minute telephone interview might be more exhausting than a 15-minute face-to-face interview, there being more social and non-verbal cues in face-to-face interaction. On the other hand in outsourced telephone interviews care has to be taken on standardization of the conduct of the interview, the content of questions, the entry of responses and, indeed, to check that the interviews have been done and response not simply fabricated.

In conducting telephone interviews it is important to consider several issues, for example:

1   Will the people have the information that you require? Who will you need to speak to on the telephone? If the person answering the call is not the most suitable person then you need to talk to somebody else.

2   There is a need to pilot the interview schedule and to prepare and train the telephonists, and to discover the difficult/sensitive/annoying/personal questions, the questions over which the respondents hesitate and answer very easily; the questions that will need prompts and explanations.

3   Keep to the same, simple response categories for several questions, so that the respondents become used to these and keep in the same mindset for responding.

4   Keep personal details, if any, until the end of the interview, in order to reduce a sense of threat.

5   Keep to no more than, at the most, 35 questions, and to no more than, at the most, 15 minutes, and preferably ten minutes.

6   Clear with the respondents at the start of the interview that they have the time to answer and that they have the information sought (i.e. that they are suitable respondents). If they are not the most suitable respondents then ask if there is someone present on the premises who can answer the questions, or try to arrange callback times when the most suitable person can be reached. Ask to speak to the most suitable person.

7   Keep the terminology simple and to the point, avoiding jargon and confusion.

8   You should be able to tell the gender of the respondent by his or her voice, i.e. there may be no need to ask a particular question.

9   Keep the response categories very simple and use them consistently (e.g. a mark out of ten, ‘strongly agree’ to ‘strongly disagree’, a 1–5 scale, etc.);

10   Rather than asking direct personal questions (unless you are confident of an answer), e.g. about age, income, ask about groups (e.g. which age group do they fall into (and give the age groups) or income brackets (and give them)).

Telephone interviewing is a useful, but tricky art.

We sum up the different forms of administering interviews in Figure 21.1.

FIGURE 21.1 Methods of administering interviews

21.12 Ethical issues in interviewing

Interviews have an ethical dimension; they concern interpersonal interaction and produce information about the human condition. Though one can identify three main areas of ethical issues here – informed consent, confidentiality and the consequences of the interviews – these need to be ‘unpacked’ a little, as each is not unproblematic (Kvale, 1996: 111–20). For instance, who should give the informed consent (e.g. participants, their superiors), and for whom and what? How much information should be given, and to whom? What is legitimate private and public knowledge? How might the research help or harm the interviewees? Does the interviewer have a duty to point out the possible harmful consequences of the research data or will this illegitimately steer the interview?

It is difficult to lay down hard and fast ethical rules, as, by definition, ethical matters are contestable. Nevertheless it is possible to raise some ethical questions to which answers need to be given before the interviews commence:

   Has the informed consent of the interviewees been gained?

   Has this been obtained in writing or orally?

   How much information should be given in advance of the study?

   How can adequate information be provided if the study is exploratory?

   Have the possible consequences of the research been made clear to the participants?

   Has care been taken to prevent any harmful effects of the research to the participants (and to others)?

   To what extent do any potential benefits outweigh the potential harm done by the research, and how justifiable is this for conducting the research?

   How will the research benefit the participants?

   Who will benefit from the research?

   To what extent is there reciprocity between what participants give to and receive from the research?

   Have confidentiality, anonymity, non-identifiability and non-traceability been guaranteed? Should participants’ identities be disguised?

   How do Data Protection Acts and laws operate in interview situations?

   Who will have access to the data?

   What has been done to ensure that the interview is conducted in an appropriate, non-stressful, non-threatening, manner?

   How will the data and transcriptions be verified, and by whom?

   Who will see the results of the research? Will some parts be withheld? Who own the data? At what stage does ownership of the data pass from interviewees to interviewers? Are there rights of veto for what appears? To whom should sensitive data be made available (e.g. should interview data on child abuse or drug taking be made available with or without consent to parents and the police)?

   How far should the researcher’s own agenda and views predominate? What if the researcher makes a different interpretation from the interviewee? Should the interviewees be told, even if they have not asked for these interpretations?

These issues, by no means an exhaustive list, are not exclusive to the research interview, though they are highly applicable here. For further reading on ethical issues we refer readers to Chapter 5. The personal safety of interviewers must also be addressed: it may be important, for example, for the interviewer to be accompanied, to leave details of where he or she is going, to take a friend, to show identification, to take a mobile phone, to reconnoitre the neighbourhood, to learn how to behave with fierce dogs, to use the most suitable transport. It is perhaps a sad indictment on society that these considerations have to be addressed, but they do.

For further information on interviewing we refer the reader to the following websites:

www.bera.ac.uk/data-collection-interviews-in-research/ http://fds.oup.com/www.oup.co.uk/

pdf/0–19–874204–5chapl5.pdf

www.edu.plymouth,ac.uk/resined/interviews/inthome.htm

http://libraries.uta.edu/dillard/subfiles/SocWSocSciInter-viewingBib.htm

http://pareonline.net/getvn.asp?v=5&n=12

www.eric.ed.gov/ERICDocs/data/ericdocs2sqFcontent_storage_01/0000019b/80/27/f2/8d.pdf

www.utexas.edu/academic/diia/assessment/iar/research/plan/method/interview.php

www.public.asu.edu/~ifmls/artinculturalcontextsfolder/qualintermeth.html

www500/Interview%20Fri.pdf

www.sasked.gov.sk.ca/docs/social/psych3 0/support_materials/research_methods.htm

www.sociology.org.uk/methfi.pdf

www.gerardkeegan.co.uk/resource/interviewmeth1.htm

http://owl.english.purdue.edu/owl/resource/559/06/

http://owl.english.purdue.edu/owl/resource/559/04/

www.vanuatu.usp.ac.fj/student_resources/Resources_Main/interviews.html

http://managementhelp.org/evaluatn/intrview.htm

http://ppa.aces.uiuc.edu/KeyInform.htm

www.scu.edu.au/schools/gcm/ar/arp/iview.html

Companion Website

The companion website to the book includes PowerPoint slides for this chapter, which list the structure of the chapter and then provide a summary of the key points in each of its sections. In addition there is further information in the form of an example of interviewing and interview transcripts. These resources can be found online at www.routledge.com/textbooks/cohen7e.