Social Research Glossary A B C D E F G H I J K L M N O P Q R S T U V W X Y Z Home
Citation reference: Harvey, L., 2012-24, Social Research Glossary, Quality Research International, http://www.qualityresearchinternational.com/socialresearch/
This is a dynamic glossary and the author would welcome any e-mail suggestions for additions or amendments.
|
|
_________________________________________________________________
Interview
An interview is a method of collecting data from a subject by asking questions in a face-to-face situation.
Introduction
There are several different types of interview.
Types of Interviews
A structured (or standardised) interview is one in which every effort is made to exclude procedural reactivity by asking every subject exactly the same question in the same way and in the same order. The emphasis is upon `equivalence of stimulus'.
The interview will be based upon an interview schedule which is a pre-set list of questions and instructions used by the interviewer in the field with the aim of standardizing the interview procedure. The interview schedule imposes structure on data collection. The schedule is sometimes mis-called a questionnaire.
Generally, the interview procedures and schedule will be of the outcome of considerable pilot work.
An exploratory interview is unstructured and is intended to develop ideas and reseach hypotheses, and to explore possible ways of gathering relevant data. Often it is based upon a checklist of items to be covered in the interview.
Not all unstructured interviews are exploratory interviews in preparation for structured interviews. They may constitute an ethnographic method in their own right.
A pilot interview is intended not for data collection as such but as an aid to the design of later research. Early pilot interviews will generally be unstructured, probing for subjects, language and concepts, but later ones may test the suitability of standardized instruments.
Pilot interviews are of two sorts. The first is 'exploratory' and is an initial attempt to see if the subject being investigated is adequately (or potentially adequately) captured by the proposed interview procedure and schedule of questions. This is more aptly called a pre-pilot.
Second, is the complete pilot, which is a small-scale implementation of the complete research plan including data collection, analysis and outline report. The idea is to see whether the full-scale research that is proposed will actually investigate what it is intended to. Problems with wording of questions, interview techniques, coding, data analysis and the matching of intentions with available data are sorted out at this stage before considerable resourses are committed for the full-scale research.
An in-depth interview (or depth interview) is one in which the researcher tries to go deeply into some aspect of the respondent's feelings, motives, attitudes, life history, etc.
The intention here is to get respondents talking in detail about a particular aspect of their experience and their reflections on it. The direction and nature of the responses is directed, at least initially, by the respondent. The structure of the interview may be one in which a very general topic is gradually narrowed down. For example, respondents may be asked to talk about their life in general and gradually focus down onto their work experiences. Or the topic may be more specific to start with and then broadened out to. For example, the start point might be the respondent’s current work, and the interview gradually broadened to address prior work experience, relations between workers and management, issues of gender and race at work, and so on.
The ostensive focus of the enquiry may never be made clear to the respondent because the researcher may be interested in the ways that the respondent naturally endorses or assimilates particular attitudes or perspectives or engages in particular behaviour. For example, a researcher may be interested in attitudes to class but never address this directly, instead seeing how comments on life and work experiences are class related. Or a researcher may be interested in the extent to which football supporters have been active in ‘hooligan’ activity, but rather than direct attention to this, get the respondent to talk about their enjoyment of football, whether they go to many matches, and where.
In order to get to the respondent’s meanings, the depth interviewer has to be alert to underlying nuances and seek out the taken-for-granteds of the respondent. This requires gentle probing and the ability of interviewers to ‘think on their feet’. Where the possibility arises, it is advisable to have at least one follow-up to an in-depth interview to pick up themes from the first interview once the transcript or recording of the first interview has been examined.
In the main, depth interviews are used as part of an ethnographic approach. However, in-depth interviews are sometimes used to develop sensitizing concepts, to get at the respondent’s everyday categorisations and typifications, or to gain some feeling for the respondents’ language style before designing a questionnaire or scheduled interview.
See Researching the Real World section 4.1
Semi-structured interviews are as the name suggests based on an outline structure or some key questions but with a degree of latitude for the interviewer to explore the topic in more detail, depending on how the conversation is going. Semi-structured interview is usually one in which the interviewer has a check list of questions that the respondent is asked to address.
The intention is to get respondents to talk in their own terms, hence questions tend not to be too specific. If in the course of talking about one area, the respondent provides answers to another area then the interviewer checks this off. The questions do not need to be asked in any given order, rather they should be asked in a way that develops the conversation. It is much less rigid than a strcutured interview and permits the interviewer to provide information as well as receive it. Semi-structured interviewing often starts with more general questions or topics and arguably many questions emerge during the interview, within a pre-set general framework.
The extent to which the semi-structured interview develops the frame of reference of the respondent, rather than getting detailed replies to questions that are pre-designed by the researcher, depends on the extent to which the interviewer is encouraged, or is able, to probe replies in order to draw out respondents' meanings in their own terms. If probes are pre-defined, as in structured interviewing, then the semi-structured interview is less likely to go beyond the researcher's categories than if the interviewer is free to pursue an independent line of probing designed to seek out the individual respondent's perspectives and frame of reference.
Unstructured interview (sometimes referred to as 'open-ended interview')
Unstructured interviews have, as their name suggests, no formal structure and are more like everyday conversations. There will be a topic or set of topics, which may be quite broad, that the researcher wants to discuss in the interview but there will be no set questions. Unstructured interviews, unlike structured interviews, are unconcerned with asking respondents the same questions in the same order. The aim of unstructured interviewing is to access the point of view of the respondent using the respondent's frame of reference rather than the researcher's pre-structured frame. Unstructured interviews thus usually set out to uncover the meanings that the respondent constructs about aspects in the social world.
The conversation will be fairly free-flowing and the questions of probes should build upon what the respondent says, even if at times they seem to be straying off the topic. This is because the interviewer may have artificially restricted the topic in his/her mind and the openness of the unstructured interview permits exploration of areas not formerly considered. However, the interviewer may need to direct the conversation back to the main topic if the discussion seems to be going off at a tangent. Unstructured interviews are far more difficult to conduct than structured interviews as the interviewer usually needs to be veyr well versed with the underlying research thesis, needs to stay alert to clues and nuances and be prepared to respond to questions as well as ask them.
Unstructured interviewing is more likely to be used as part of an ethnographic approach than in conjunction with quantitative techniques. However, unstructured interviews are sometimes used as exploratory interviews to determine the content of a structured interview schedule.
Unstructured interviews range from casual conversations, or chats, through in-depth interviews and ethnographic interviews to semi-structured interviews. These different forms of unstructured interviews tend to overlap and do not have discrete meanings.
In all unstructured interviewing the respondent's replies are, ideally, recorded verbatim, through short-hand or mechanically on some form of recording media.
Dialogic interview is a general term for interviews that argue for a true dialogue and in so doing adopt an approach that involves a radical reappraisals of the conventional view of interviews as one-way data gathering tools.
Dialogic interviews are similar to unstructured interviews in as much as the interaction between interviewer and interviewee is a conversation rather than a one-way interrogation. Rather more than unstructured interviews, though, dialogic interviews allow the interviewee to control the interview. That is, the conversation goes in the direction that the interviewee wants to go.
This is a technique, for example, used in some counselling or psychoanalytic sessions. In the social sciences, this approach tends to be mainly used in ethnographic work, and is often part of a broad participant-observation approach to the research. As with unstructured interviews the dialogic interview does not require standardised question but it does require that the interviewee understands the underlying purpose and thesis of the research endeavour.
The conventional view is that data is collected best through interviews where the interviewer asks questions in a relaxed and conversational way as possible without actually engaging in a two-way conversation and simply records answers with as little comment as possible while encouraging the respondent by providing affirmative signs, such as the occasional nod, etc. It has been argued, on the contrary, that there is a contradiction between the development of rapport and the unidirectional and inherent power-relationship involved in seeking out information that the researcher wants.
This critique has been most fully developed in feminist analyses of the standard interview relationship (and dialogic interviews are sometimes referred to as feminist interviews). This conventional approach is seen as embodying a male-paradigm of enquiry. This paradigm prefers a mode of enquiry in which the respondent is encouraged into an attitude in which they give information to a interviewer. The apparent dialogue between interviewer and respondent, dressed up as rapport, is seen as entirely spurious. The flow of information is one way, the interviewer giving nothing back. The lack of reciprocity is linked to the scientistic idea of objectivity manifested in the interviewers ‘disinterestedness’. This is seen as a male paradigm in distinction to the sharing of experiences that is characteristic of sisterhood.
The alternative model proposed is one in which a real dialogue takes place. Interviewer and respondent take an equal share in the process and information is exchanged. The interview becomes of mutual benefit to all participants and the power-relationship of conventional interviewing is broken down. rather than seeing this as subjective, proponents of this model argue that far more insight is obtained, not only into the respondent’s perspectives,
but also the interviewer’s taken-for-granteds.
Furthermore, in the process of a mutually beneficial exchange, a political intent is also inaugurated, the reflection of both parties on the nature of the oppressive structure in which they operate.
In general, the dialogic interview is intrinsic to critical ethnography.
Ethnographic interview refers to any kind of unstructured interview which is directed to ethnographic ends; viz. the determination of the respondent's meanings.
Usually, ethnographic interviewing consists of a series of interviews, which may be a sequence of in-depth interviews; an ever more specific set of interviews ranging from conversations through directed depth-interviews to a semi-structured interview.
Ethnographic interviewing should encourage the respondent to talk in their own terms and gradually probe the meanings of terms and frames of reference. Initially, the researcher begins to understand terminology and perspectives through their context in responses. Over time, the concepts and relationships that appear to be important in the way respondent's construct meaning can bedirectly engaged by asking respondent's to explain them in their terms.
Ethnographic interview is a term most often used to refer to a formal interview of one sort or another, in contradistinction to participatory conversations, or over heard conversations which take place in a participant observation or non-participant observation context. However, some commentators include the 'interviewing' undertaken by an observer in the category of ethnographic interviews.
The ethnographic interview, in general, is distinguished from a conversation by its formality. The ethnographic interview will be more formal (although conventionally this does not mean interrogatory or lacking in rapport) than a conversation because the interviewer must make clear the purpose of the interview and get the respondent's agreement to the recording of the exchange. The formality may become more structured in later stages in the ethnographic interview (series) if the respondent is asked to perform tasks, such as order a set of items, provide documents, draw maps.
Bias in interviews
Interviewer bias/interviewer effect
In all types of interview one runs the danger of interviewer bias which occurs when a particular observation or response is influenced by some attribute of the interviewer. This may be a result, for example, of the way in which interviewers presents themselves, the impression they give, the way in which they interpret answers, and the way in which they 'lead' the respondent. Interviewer bias is one of several sources of bias in social surveys. It operates in a very subtle manner and is often difficult to detect and measure and yet can affect the results considerably.
Qualitative interview is a term used to refer to an array of interview types that seek to collect 'qualitative' as opposed to 'quantitative' results. This is a rather simplistic dichotimisation. (see the exerpt from Edwards and Holland (2013) below)
Interview effect refers to the impact that the artificial situation of the interview has on the information respondents are prepared to provide, their attitudes and opinions.
Despite attempts at rapport, most interviews (and all structured ones) are contrived one way exchanges which abrogate normal conversational practices. The resultant information is in no way naturalistic. Information exchanged is thus likely to be highly selective and often, in practice, that part of the information that the respondent thinks the interviewer wants to hear and that the respondent is prepared to make available.
Ethnographic interviews in part overcome this by repeated interviews, adopting more flexible approaches and by exploring a number of apparently peripheral lines of enquiry. Such interviews tend to be lengthy.
Hired-hand effect refers to the bias, errors or distortion that can creep into data as a result of using employees who are less committed to the enquiry than the researcher.
Hired-hand effect can vary from the accidental mischecking of an item on a schedule, to the failure of adequately recoding open (verbatim) answers, to the inability to follow the appropriate 'route' through a schedule, to deliberately questionning the wrong people out of convenience, or even, outright fraudulent completion of schedules by the interviewer without interviewing anyone. Such effects, especially minor accidental mistakes, are often difficult to determine.
Cultural incompatibility occurs when interviewer and respondent have very different socio-cultural backgrounds or political views such that each does not fully (or at all) understand what the other is saying.
Santiago (2009) wrote :
There are three types of interviews that can be conducted when carrying out a research project. Those are structured, semi-structured, and unstructured. This article will describe what each one involves and the differences between them.
Structured Interviews: Structured interviews require adherence to a very particular set of rules. Each question that is outlined should be read word for word by the researcher without any deviation from the protocol. In some cases, the interviewer is also required to show consistency in behavior across all interviews. This includes bodily posture, facial expressions, and emotional affect. Reactions to participant responses should be kept to a minimum or avoided entirely.
Structured interviews are the type used most often by quantitative researchers. The style is most useful when looking for very specific information. The benefits are that it keeps the data concise and reduces researcher bias.
Semi-structured Interviews:
Semi-structured interviews are a bit more relaxed than structured interviews. While researchers using this type are still expected to cover every question in the protocol, they have some wiggle room to explore participant responses by asking for clarification or additional information. Interviewers also have the freedom to be more friendly and sociable.
Semi-structured interviews are most often used in qualitative studies. The style is most useful when one is investigating a topic that is very personal to participants. Benefits include the ability to gain rapport and participants' trust, as well as a deeper understanding of responses. Data sets obtained using this style will larger than those with structured interviews.
Unstructured Interviews:
Unstructured interviews have the most relaxed rules of the three. In this type, researchers need only a checklist of topics to be covered during the interview. There is no order and no script. The interaction between the participant and the researcher is more like a conversation than an interview.
Unstructured interviews are most often used in ethnographies and case studies (types of qualitative studies). They are best used when researchers want to find as much information as possible about their topic. The benefit is that unstructured interviews often uncover information that would not have been exposed using structured or semi-structured interviews. The researcher and participant are not limited by the protocol. Data sets collected using unstructured interviews will be larger than the rest.
Colorado State University (1993–2013) defines interview as
Interviews: A research tool in which a researcher asks questions of participants; interviews are often audio- or video-taped for later transcription and analysis.
Interview|: A face-to-face or telephone questioning of a respondent to obtain desired information perspective:.
Polit and Beck (2006) define an interview as:
A method of data collection in which one person (an interviewer) asks questions of another person (a respondent): interviews are conducted either face-to-face or by telephone.
Mason (2004) defined a structured interview as follows:
Semistructured interviewing is an overarching term used to describe a range of different forms of interviewing most commonly associated with qualitative research. The defining characteristic of semistructured interviews is that they have a flexible and fluid structure, unlike structured interviews, which contain a structured sequence of questions to be asked in the same way of all interviewees. The structure of a semistructured interview is usually organized around an aide memoire or interview guide. This contains topics, themes, or areas to be covered during the course of the interview, rather than a sequenced script of standardized questions. The aim is usually to ensure flexibility in how and in what sequence questions are asked, and in whether and how particular areas might be followed up and developed with different interviewees....
Sociology Central (undated) describes what it calls 'focused (semi-structured) interviews
This technique is used to collect qualitative data by setting up a situation (the interview) that allows a respondent the time and scope to talk about their opinions on a particular subject. The focus of the interview is decided by the researcher and there may be areas the researcher is interested in exploring.
The objective is to understand the respondent's point of view rather than make generalisations about behaviour. It uses open-ended questions, some suggested by the researcher (“Tell me about...”) and some arise naturally during the interview (“You said a moment ago...can you tell me more?”).
The researcher tries to build a rapport with the respondent and the interview is like a conversation. Questions are asked when the interviewer feels it is appropriate to ask them. They may be prepared questions or questions that occur to the researcher during the interview. The wording of questions will not necessarily be the same for all respondents.
Bryant (undated) outlines unstructured interviews
Unstructured interviews are the opposite to structured interviews. Unstructured interviews are more like an everyday conversation. They tend to be more informal, open ended, flexible and free flowing. Questions are not pre-set, although there are usually certain topics that the researchers wish to cover. This gives the interview some structure and direction. An unstructured interview is “an interview without any set format but in which the interviewer may have some key questions formulated in advance. Unstructured interviews allow questions based on an interviewee's responses and proceeds like a friendly, non-threatening conversation. However, because each interviewee is asked a different series of questions, this style can lack the reliability and precision of a structured interview. Unstructured interviews are also called non-directive interview.” [the quote within this extract is unattributed]
Edwards and Holland (2013, pp. 2–4)
Most text books will tell you that interviews range through a continuum, from structured, through semi-structured, to unstructured (or focused) interviews (Bryman 2001, May 1997). The structured interview is at the quantitative end of the scale, and more used in survey approaches. The rest of the scale, semi-structured and unstructured, is the area occupied by qualitative researchers, with the interviews characterized by increasing levels of flexibility and lack of structure. Many of the terms you will have discovered applied to qualitative interviewing appear in this part of the continuum, for example in-depth, informal, non-directed, open-ended, conversational, naturalistic, narrative, biographical, oral or life history, eth- nographic and many more .... The terms used for any particular interview type relate to the underlying philosophy and specific approach taken to research....
Briefly, the structured interview is based on a questionnaire with a sequence of questions, asked in the same order and the same way of all subjects of the research, with little flexibility available to the researcher. The major objective is for neutral interviewers to obtain comparable infor- mation from a potentially large number of subjects. It is typical of more positivist approaches, with methodological rules for its practice, and often is subjected to statistical methods of analysis.
A considerable range of qualitative approaches use semi-structured and unstructured interviews, as suggested by the list above. Jennifer Mason argues that, despite the large variations in style and tradition, all qualitative and semi-structured interviewing has certain core features in common:
1. The interactional exchange of dialogue (between two or more participants, in face-to-face or other contexts).
2. A thematic, topic-centred, biographical or narrative approach where the researcher has topics, themes or issues they wish to cover, but with a fluid and flexible structure.
3. A perspective regarding knowledge as situated and contextual, requiring the researcher to ensure that relevant contexts are brought into focus so that the situated knowledge can be produced. Meanings and understandings are created in an interaction, which is effectively a co-production, involving the construction or reconstruction of knowledge. [Adapted from Mason 2002: 62]
With a colleague, one of us has pointed out that as well as being interac- tional, the interview is a social and potentially a learning event for both participants:
As a social event it has its own set of interactional rules which may be more or less explicit, more or less recognized by the participants. In addition . . . there are several ways in which the interviewconstitutes a learning process. . . . Participants can discover, uncover or generate the rules by which they are playing this particular game. The interviewer can become more adept at interviewing, in terms of the strategies which are appropriate for eliciting responses. (Holland and Ramazanoglu 1994: 135)
Both interviewers and interviewees can learn more about certain aspects of themselves and the other, with or without this being an explicit part of the interactional exchange.
The interviewer's job consists of:
a) locating the respondent
b) persuading them to answer questions
c) asking questions
d) recording answers
e) ensuring answers are 'meaningful'
f) ensuring the answers are the respondents’ own.
The United Kingdom Government Social Survey Division developed a standardised procedure for interviewers to follow, and this agency has probably been the most scrupulous and efficient scheduled interviewing body in the country. In her Handbook for Interviewers, Jean Atkinson (1964) outlined the procedure for interviewing for employees of the Government Social Survey Division. This approach will be referred to here as the 'official method'.
Most interviewers in the Social Survey Division in the 1970s were middle-class white women. Their training encouraged them to be polite and pleasant but with a slightly superior air of efficiency. They were encouraged to dress 'smartly' in, what official circles would regard as, 'neutral' dress. All in all they had an indubitably 'official' air about them.
Thus handicapped the Government Social Surveys Division interviewers set forth, usually in search of members of a randomly selected sample. In other words, they went to an address and attempted to interview a named sample member. A good interviewer is not easily discouraged. If there is no one in, she goes back later. If she fails to contact respondent she may even go next door to find out if the neighbour knows where the respondent is.
Once the respondent is located, the interviewer will arrange a suitable time for the interview, (which might be immediately), will explain the purpose and indicate sponsorship of the survey, and point out the importance of recording the selected sample members' response. (To avoid bias in the random sample). The interviewer (at whatever appears to be a tactful time) will point out the confidentiality of the survey and hence of the respondents' replies; (confidentiality does not mean anonymity, however). When arranging the time of interview, the respondent is told approximately how long the interview will last. This ensures that the respondent does not feel 'put upon' when the interview drags on for more than ten minutes.
Once the respondent has agreed to the interview; the interviewer should take gentle command of the situation. On entry to the residence, (no government surveys are really short enough to be conducted on the doorstep, nor does the official approach recommend this), the interviewer should, as far as possible, isolate the respondent from his/her family and get him/her into a comfortable position, (usually seated facing the interviewer who would normally also be seated). This prevents the respondent from reading the schedule over the interviewer's shoulder and relaxes the respondent. Then the questions begin.
Questions should be read in as conversational a tone as possible. Questions should flow, should not be read in a monotone, and the interviewer should appear to be interested in the questions and enthusiastic about the survey (but should not overdo it). Questions should be read at a reasonable pace neither too fast nor too slow, and 'pregnant pauses' should be avoided.
The interviewer must aim at putting the respondent at ease, she aims at achieving 'rapport' with the respondent. This, it is considered, will maximise the utility of the interview situation, resulting in an easy flow of required information from the respondent. Therefore, the interviewer must be familiar with the survey and the questions and must not appear to be reading them for the first time.
The interviewer has to record the answers, and in doing so must ensure she makes no mistakes. Answers to pre-coded questions are the easiest to record, it just involves marking the appropriate code on the schedule. If the pre-coded question is of the sort that involves the respondent selecting an answer from a list then recording the answer is simple (although the respondent may select an answer as required, but also qualify it—such qualification ought to be recorded). If the pre-coded categories are not known to the respondent (i.e. are just on the interviewer's schedule) then the interviewer has to fit the response given by the interviewee to one of the pre-coded categories. Recording answers to open questions is the hardest work as it requires verbatim recording of the reply. In practice the interviewer may not be able to record every word, particularly if the respondent is unusually loquacious, because she can not ask the interviewee to slow down whilst she records the answer for fear of endangering the rapport.
The 'official methods' demands an almost ritualistic adherence to the wording of the question. Each question must, at least initially, be read out exactly as it is written. No variation is tolerated. The presentation of the question (e.g. emphasis on words etc) should not vary either. This ensures that all respondents are answering the same question and the responses can be meaningfully compared. As far as opinion questions go, no variation is permitted. If no opinion is offered, the question may be (tactfully) repeated; but that is all. If an opinion is expressed, which is either ambivalent or cliched, then the interviewer is required to probe. Such probes are of two sorts: clarifying probes to clear up ambivalence in verbatim recorded answers to 'open' questions, or to clarify the meaning so that the response may be coded; and exploratory probes that are used to check the depth of opinion or understanding of the question to which an opinion has been stated. (Schuman (1966) suggests the use of 'random probes' on pre-coded questions in order to assess understanding). An example of such a probe is "can you explain a little more fully what you mean by that". The 'official approach' specifies the wording of such probes as may be used, but as Schuman points out it is the actual delivery, not the wording of the probe that is important. Interviewers should never appear to be threatening or challenging if rapport is to be maintained.
Factual questions are allowed a little more leeway than opinion questions. Initially each respondent is asked the same question. If it is not understood fully by the respondent, the interviewer is allowed to rephase it in order to get at the 'facts'. If the facts seem doubtful or are not forthcoming the interviewer is allowed to restate the question or use standard prompts to dig the facts out. If a response appears to be in error or is unlikely to be true, the interviewer may use the standard probes to illicit the truth. However, the interviewer must avoid turning the interview into a third-degree interrogation.
The 'official method' tries hard to ensure comparability of responses ensuring each respondent is asked the same questions in the same order. The responses are checked for 'truth' and 'understanding'. The interviewer employed by the Social Survey Division are good at overcoming facetiousness, exaggeration and downright lying in a polite and firm way. (But even they are not always successful, of course). Whilst appearing enthusiastic, the interviewers must also appear 'neutral'. If they have opinions on the subject matter of the survey they should not become explicit, and certainly not cause the interviewer to emphasise certain words, thus making questions into 'leading; ones; nor cause the interviewer to use biased prompts or suggest responses when interviewees are unsure or unable to fully express what they wish to say. Any hints by an interviewer, deliberate or otherwise, will be easily picked up by a respondent who is almost certainly trying to guess what the interviewer thinks about the survey subject. Interviewers should also avoid anticipating the response. The socio-economic status, sex or age of a respondent might lead the interviewer to expect a particular response , resulting in a faulty record of the reply because the interviewer does not pay full attention to what the respondent is saying.
The interviewing model and studies of bias in interviewing
The Government Social Survey Division is probably the most conscientious survey organisation, and tries hard to overcome the obvious difficulties that arise in the interview situation. They represent a model of scheduled interview surveying that other agencies approach. There is no doubt that however good the interviewing is, bias (in the conventional sense) still does creep in and it is, of course, worse in less scrupulous agencies who are unaware of, or don't care about, the possible sources of bias.
Deming (1944) noted that interviewer bias was a well-established source of error in surveys. He reports a study by Stuart A Rice (undertaken in 1914 and reported in American Journal of Sociology 1929) that the response to an interview can be highly influenced by the interviewer. Two thousand 'destitute men' were interviewed: their destitution was attributed to 'the evils of alcohol' by a prohibitionist and to 'industrialisation' by a socialist.
Deming suggests that the education, economic status, environmental background, political, religious and social beliefs of interviewers will all affect the interview situation and thus the results. He said: 'One source of bias and variability arising from the interviewer has its roots in lack of understanding of the subject and purpose under investigation, without which the interviewer cannot evaluate a situation or properly record the respondent's statements' (Deming, [1944] 1970, p. 327)
J. Allen Williams (1964) showed what effect 'social distance' had on response variation. The study was undertaken in the United States and 'social distance' was measured on a 'race' and 'socio-economic class' continuum. The greatest social distance being between upper-middle class white interviewers and lower-class black respondent. All respondents were black, interviewers were of both races. The largest potential bias occurred when questions were 'threatening' that is, they challenged accepted social norms. In particular, Williams noted, that middle-class white interviewers obtained a significantly greater percentage of conservative responses from lower-status blacks than did middle-class black interviewers on these highly threatening questions. While this clearly illustrates the likely bias due to the differences in socio-economic status of the interviewer and the interviewee it is an extreme case. Nonetheless, it would suggest that social differences will affect replies. In a reassessment of this earlier work and an extension of the research, Williams (1968) stated 'it seems likely that the role performance of the interviewer could either enhance or mitigate the biasing effects of status characteristics and potentially threatening questions'. The 'objectivity' of the interviewer and degree of 'rapport' are, he noted, the two main factors contributing to a successful interview, a view supported by other researchers (Kornhauser & Sheatsley, 1959; Richardson et al., 1965; Kahn & Cannell, 1958).
However, other researchers have pointed out that too much rapport may cause a strain towards consensus either consciously or subconsciously (Newcombe, 1961; Emery et al., 1957; Thibaut & Kelley, 1959)). Williams' follow up research rather amplified the problems of achieving a successful interview, and he concluded that 'the crucial question that remains unanswered is which combination of factors minimise bias' (Williams, [1968] 1970, p. 231)
Some researchers have cast doubts on the reliability and validity of scheduled interview data (Hyman, 1944; Gurman and Bass, 1961; Nye and Short, 1957; Roth, 1970; Chein, 1964; Blum, et al., 1964; Scott & Willcox, 1965; Gillies, 1965; Clark & Tifft, 1966; Maccoby and Maccoby, 1954). Hagbury ([1968], 1970, p. 279), for example, noted, in his research on a group of participants in an adult education programme, that 'in general...(the) responses are in the direction of ideal norms.' Ball ([1967] 1970, p. 236), in analysing deviant behaviour, took his analysis a little further to suggest that 'the social situation and auspices under which the interviews are obtained affect the deviant subjects motivation to be either candid, equivocal, or deceitful'. His study of narcotic drug addicts (a field commonly believed to be unnassailable by scheduled interviewing) clearly shows that reliable results may be obtained but that familiarity with the subculture and slum conditions, the lack of any police or social service function, and experienced interviewing were essential features in the relative success he had.
In summary, then, it appears that two main points emerge as of prime importance for reducing bias in scheduled interviews, social distance and auspices. Social distance does not simply involve superficially distant status between interviewer and interviewee, awareness of the interviewee and her/his other environment is more important, not just for 'rapport' but in order to improve the interviewers understanding of the respondent's answers.
The auspices has a large bearing. In general, deviant groups would seem to be more forthcoming if no 'official' label is attached to the research; whilst a survey that is not aimed at deviant groups will probably achieve more if it has an official and 'high falluting' backing. There are, of course, exceptions to this generalisation. However, no matter how well the interview is carried out, it would be a very brave or naive researcher who would ignore the possible bias in the results, even from a 'perfectly' worded schedule.
Without doubt the interview process incorporates various biasing effects that are normally referred to as 'interview bias' without being differentiated into its component parts. The 'official method' tries hard to overcome this bias but it can only be partly successful; a fact that will become more obvious if one considers the total likely causes of bias in the interview situation and break it down into its component parts.
The study by Ball ([1967] 1970) cited above suggested that one of the main reasons for his relative success was 'the interviewer's knowledge of the addict subculture and familiarity with lower-class slum neighbourhoods.' This is a clear indication of the first area of bias in interviews namely cultural incompatibility of interviewer and respondent. A second source of bias in interviews is interviewer effect: the interviewee gives the replies he or she thinks ould impress the interviewer. A third aspect is that which results almost inevitably in any scheduled interview survey, namely the bias that arises using paid employees to do the interviewing. The biasing effects due to disinterested, harassed, over-worked or non 'objective' employees is known as the hired-hand effect (Roth, 1966). Finally, the very uniqueness of the interview situations results in interview reaction. The invasion of the respondents' routine, usually by a complete stranger, will cause him or her to adapt a new role and must affect behaviour with inevitable biasing effects. The extent and direction of these four different biases can never really be assessed. (With luck they may cancel each other out, but to expect such a result would be wishful thinking).
The 'official' approach to interviewing, whilst attempting to reduce interviewer effect and minimise hired hand effect, does very little to alleviate cultural incompatibility and, if anything, accentuates interview reaction (interview effect).
John C. Ball, who achieved success when using a structured interview with narcotic drug addicts did so primarily because he overcame many of these problem areas reducing potential bias. He used a very competent interviewer, who was able to overcame the problems of cultural incompatibility to a large extent, and as such Ball's study was not prone to the bias of hired-hand effect. The interview reaction/effect in Ball's study was somewhat reduced by prior contact with the respondents and the interviewer effect by the research's lack of official auspices. In this sense Ball's study was rather unusual for a scheduled interview survey and larger scale, less intimate approaches are the norm.
See also
Researching the Real World section 4.4.3
Researching the Real World Section 8.3.10.2
Atkinson, J., 1964, Handbook for Interviewers. London, HMSO.
Ball, J.C., 1967, 'The Reliability and Validity of Interview Data Obtained from 59 Narcotic Drug Addicts', American Journal of Sociology , 72, pp. 650–54. Reprinted in Forcese & Richer (Eds.), 1970, Stages of Social Resarch, Englewood Cliffs, Prentice-Hall pp. 231–36.
Blum, R. & Associates, 1964, Utopiate . New York, Atherton Press.
Bryant, L., undated, 'Unstructured interviews' on History Learning Site (2002–2013) available at http://www.historylearningsite.co.uk/unstructured_interviews.htm, accessed 6 March 2013, still available 7 June 2019 (now dated 20 May 2019 with annoying 'recommended cookies' overlay).
Chein, I., et al., 1964, The Road to H. New York, Basic Books.
Clark, J. & Tifft, L., 1966, 'Polygraph and interview validation of self reported deviant behaviour', American Sociological Review, 4, pp. 516–23.
Deming, W.E., 1944, 'On errors in surveys', American Scological Review, 19, pp. 359–69. Reprinted in Dezin (ed), 1970, Soclogical Methods : A Sourcebook, London, Butterworth, pp. 320–37.
Edwards. R. and Holland, J., 2013. What is Qualitative Interviewing? London, Bloomsbury, available at http://eprints.ncrm.ac.uk/3276/1/complete_proofs.pdf, accessed 9 March 2014, still available 7 June 2019.
Emery et al., 1957, Information, Decision & Action. Melbourne, Melbourne Unversity Press.
Gillies, H., 965, 'Murder in the West of Scotland' British Journal of Psychiatry, 480, pp. 1087–94.
Gurman, E.B. & Bass, B.M., 1961, 'Objective compared with subjective measures of the same behaviour in groups', Journal of Abnormal & Social Psychology, 1961, pp. 557–59.
Hagburg, E., 1968, 'Validity of questionnaire data: reported and observed attendance in an adult education program', Public Opinion Quarterly, 32, pp. 453–56. Reprinted in Forcese & Richer (Eds.), 1970, Stages of Social Resarch, Englewood Cliffs, Prentice-Hall pp. 236–39.
Hyman, H., 1944, 'Do they tell the truth?' Public Opinion Quarterly, 4, pp. 557–59.
Kahn and Cannell, 1958, The Dynamics of lnterviewing . New York, Wiley.
Kornhauser & Sheatsley, 1959, in Sellitz et al., Research Methods in Social Relations, New York, Holt, pp. 546–87.
Maccoby, E. & Maccoby, N., 1954, 'The interview: a tool of social science' in Gardner Lindzey (Ed.), 1954, Handbook of Social Psychology, Vol. 1. Reading, Mass., Addison-Wesley. Newcombe, 1961, The Aquaintance Process. New York, Holt, Reinhart & Winston.
Mason, J., 2004, 'Semistructured Interview' in Lewis-Beck, M.S., Bryman, A. and Liao, T.F. (Eds.) Sage Encyclopedia of Social Science Research Methods, available at http://srmo.sagepub.com/view/the-sage-encyclopedia-of-social-science-research-methods/n909.xml, accessed 5 March 2013, page not freely available 20 December 2016.
Nye, I & Short, F.F., 1957, 'Scaling Delinquent Behaviour', American Sociological Review , 22, pp. 226–331.
Polit, D.F. and Beck C.T., 2006, Essentials of Nursing Research. Methods, Appraisal, and Utilization. Sixth edition, Philadelphia PA, Lippincott Williams and Wilkins.
Richardson et al,, 1965, Interviewing: Its forms and functions. New York, Basic Books.
Roth, J., 1966, 'Hired hand research', American Sociologist, 1, pp. 190–96. Reprinted in Dezin (Ed.), 1970, Soclogical Methods : A Sourcebook, London, Butterworth, pp. 540–57.
Schaefer, R. T., 2017, 'Glossary' in Sociology: A brief introduction, Fourth Edition, originally c. 2000, McGraw-Hill. Available at http://novellaqalive.mhhe.com/sites/0072435569/student_view0/glossary.html, site dated 2017, accessed 11 June 2017, not found 7 June 2019.
Schuman, H., 1966, 'The random probe: a technique for evaluating the validity of closed questions', American Sociological Review, pp. 218–22 .
Scott, P. & Willcox, D., 1965, 'Delinquency and the amphetamines', British Journal of Psychiatry, 478, pp. 865–75.
Sociology Central (undated) 'Sociological Research Skills: Focused (Semi-structured) Interviews' available at http://www.sociology.org.uk/methfi.pdf, accessed 5 March 2013, still available 22 December 2016, page not at this address 7 June 2019.
Thibaut & Kelley, 1959, The Social Psychology of Groups. New York, Wiley.
Williams, J.A., 1964, 'Interviewer-respondent interaction: a study of bias in the information interview', Sociometry, 27, pp. 338–52.
Willaims, J.A., 1968, 'Interview role perforance: a further note on bias in the information interview', Public Opinion Quarterly, 32, pp. 287–94. Reprinted in Forcese & Richer (Eds.), 1970, Stages of Social Resarch, Englewood Cliffs, Prentice-Hall pp. 224–31.
Top