Alvesson, M. (2012) Views on Interviews: A Skeptical Review. In: Interpreting Interviews. London: SAGE Publications.

Structure: Interviews are divided up in terms of structured, semi-structured and unstructured. Structure is mainly a matter of degree. At one end we have the highly structured interview, where the entire process should run according to a clear plan and the delivery of interview statements should follow this neatly. At the other extreme we find almost entirely un-structured talks, where a broad theme is targeted and the researcher is open to the interview taking unexpected turns. The interviewee is free to partly define and develop the relevant sub-themes or issues, but the interviewer presumably wants to avoid too wide departures from the overall theme of the research project. In most cases, unstructured interviews are best described as loosely structured, where some questions are prepared and asked and various themes are to be covered, even if the interviewer may be fairly free and flexible also in this regard, perhaps being also inclined to follow departures from the initial area of interest. (Alvesson, 2012, pp.1-2)

Size: The size issue typically involves a choice between the single interviewee and a group of people, although interviews with pairs (e.g. couples) is a third option in some kinds of research. Within the overall area of group interviews there are different versions–from brainstorming groups with little or no structure, to highly structured settings like focus consumer groups in marketing and delphi groups where experts are tapped for their skills in forecasting, exploration and pretesting. (Alvesson, 2012, p.2)

Communication media: The meeting form (media) of interviews varies respectively between faceto-face, telephonic or electronic. The majority of the interviews of interest for qualitative research will be face-to-face. Telephonic and electronic interviews are seen as much poorer and not relevant in most cases where complex phenomena are being investigated, but are economical and can perhaps sometimes also be used for brief follow-up interviews in research aiming for ‘richer’ interview material. (Alvesson, 2012, p.2)

Category: Various groups of people are sometimes believed to call for particular considerations in terms of interview methods. Indeed, there are texts on how to interview children, old people, elites, ethnic minorities, the culturally diverse, men and women (Holstein & Gubrium, 2003b). In particular, interviews with women have attained considerable interest from feminists with attention being paid to issues surrounding power relations, ethics and the exploitation of experiences in the relationship between a researcher and the persons being studied (Olesen, 2000; Reinharz & Chase, 2003). Some would point at dominant forms of language use as being potentially problematic in relationship to certain social groups. (Alvesson, 2012, p.2)

Silverman (2006) identifies three major theoretical and epistemological positions on interviews: positivism, emotionalism and constructionism.¹ These focus respectively on facts, authentic experiences (personal meanings) and the construction of interview responses (talk). (Alvesson, 2012, p.3)

Neo-Positivism

Advocates are eager to establish a context-free truth about what is really ‘out there’ by following a research protocol and gathering responses relevant to it, minimizing researcher influence and other sources of ‘bias’. There is an interest in ‘facts’–about behaviours, practices, attitudes, values or whatever. It is viewed as vital to gain non-distorted data that can be compared or aggregated. It is therefore important that the data are not strongly affected by highly local circumstances such as the specifics of interview contexts. An interview will typically be carefully planned and tightly structured. (Alvesson, 2012, p.3)

• Not becoming involved in explanations of the study, being brief and using a standardized presentation. (Alvesson, 2012, p.3)
• Not deviating from the structure of the interview. (Alvesson, 2012, p.4)
• Being neutral and avoiding getting personal. (Alvesson, 2012, p.4)

Here, ‘the interview conversation is a pipeline for transmitting knowledge’ (Holstein & Gubrium, 1997: 113). Interview questions ‘are intended to tap individual experience’ (Charmaz, 2003: 315). Even though it is typically recognized that emphatic sensitivity and judgment are necessary and the interview cannot be conducted mechanically, this is still mainly done following the stimuli-response format. (Alvesson, 2012, p.4)

Rules, procedures, an avoidance of bias, detailed coding, large quantities of material, and so on are emphasized in methodological texts as well as in empirical writings (e.g. Charmaz, 2003; Eisenhardt, 1989; Glaser & Strauss, 1967). The ideal is viewed as a completely transparent research process, characterized by objectivity and neutrality. It is seen as vital that the researcher can clearly describe what has been done and how and that the reader should then be able to assess the work. This calls for a minimization of variation and complexity in interview work. (Alvesson, 2012, p.4)

Revised Neo-Positivism: Interactive Rationalism

The problem with this, as is being increasingly recognized, is that respondents may produce only superficial and cautious responses. Many researchers are aware of problems of trust and limited control over interviewee responses. In addition, the meaning of words used is often unclear–trying to understand meaning often calls for follow-up questions, which can sometimes create additional ambiguity as people are often incoherent in their talk (Potter & Wetherell, 1987). (Alvesson, 2012, p.4)

researchers have developed techniques such as repeat interviews in order to establish better contact, checking for consistency over time/between situations and/or giving interviewees as well as interviewers a chance to reflect upon what has been said before (e.g. Acker et al., 1991; Charmaz, 2003; Collinson, 1992). (Alvesson, 2012, p.4)

It is assumed that confidence and depth call for a fairly close interaction. It is important for a researcher to get good access to a setting, understand the language and culture of respondents, think through how to present themself and the purpose of the study, perhaps find an informant (an insider who can be helpful in guiding that researcher into the setting studied), gaining trust, establishing a rapport, etc. (Fontana & Frey, 2005). Still, this is accomplished mainly around–rather than in–interviewing, which means that an interview can be carried out in a fairly neutral manner, without social interaction leading to bias within it. (Alvesson, 2012, p.4)

Interactive rationalism to some extent recognizes social complexity and embraces ‘soft’ and flexible technical measures in dealing with the problem of how to maximize reliable responses. Techniques include continually sharing any emerging interpretations and insights with those studied, conducting interviews of various kinds and in various places (e.g. offices, homes, cafés), returning transcripts of life histories to interviewees for their feedback, performing extensive member checks by sharing interpretations with participants, talking at length with key informants, etc. (Covaleski et al., 1998: 305ff). (Alvesson, 2012, p.5)

Romanticism

Dingwall (1997) means that the nearer we come to the respondent, the closer we are to apprehending the real self. Through closeness and depth we can find the authentic and true, which are simply being expressed in our talk. Silverman (2006), while also referring to romanticism, talks about ‘emotionalism’, where the data should be about authentic subjective experiences which are revealed through unstructured, open-ended interviews. (Alvesson, 2012, p.5)

the theme of what is being addressed is not only or necessarily emotions but also ideas, values, understandings of practices, efforts to reconstruct processes and interactions (Alvesson, 2012, p.5)

The romantic researcher, advocating a more ‘genuine’ human interaction, believes in establishing a rapport, trust and commitment between interviewer and interviewee, thus turning the interview into a ‘warm’ situation. Here the interviewee is free to express him or herself authentically and will produce open, rich and trustworthy talk. These are a prerequisite in order to be able to explore the inner world (meanings, ideas, feelings, intentions) or experienced social reality of the interviewee. The typical ambition of interview-studies is to accomplish ‘deeper, fuller conceptualisations of those aspects of our subjects’ lives we are most interested in understanding’ (Miller & Glassner, 1997: 103). (Alvesson, 2012, p.5)

[understand why students are not engaging with intercultural research]

Researchers will rely on interviewees’ narrations about their lives as a way to understand them, as ‘story-telling stays closer to actual life-events than methods that elicit explanations’ (Hollway & Jefferson, 2000: 32).  (Alvesson, 2012, p.5)

Fontana and Frey (1994), for example, suggested that a researcher may reject ‘outdated’ techniques of avoiding getting involved or providing a personal opinion and instead engage in a ‘real’ conversation with ‘give and take’ and ‘emphatic understanding’ (Alvesson, 2012, p.5)

[shared my own experience and challenges]

This makes the interview more honest, morally sound, and reliable, because it treats the respondent as an equal, allows him or her to express personal feelings, and therefore presents a more ‘realistic’ picture that can be uncovered using traditional interview methods. (1994: 371) (Alvesson, 2012, p.6)

The interview subject has potentially much of value to say, but this calls for the researcher to actively lead or support that subject into intelligent talk. Interviewer and interviewee thus collaborate in the ‘co-construction of knowledge’. (Alvesson, 2012, p.6)

The major possible advantage of romantic research is instead that this kind of active approach may produce more varied and, therefore, more possible idea stimulating talk. There are no safe procedures for sorting out low-quality from trustworthy and perceptive interview accounts in terms of their descriptive value. (Alvesson, 2012, p.8)

Miller and Glassner (1997), for example, propose neutrality as a suitable response to what interviewees are saying. They see a close interaction and trustbuilding with subjects as important but believe that one should hold back one’s own opinions of interview accounts in the interaction. (Alvesson, 2012, p.8)

Kvale (1996) states that a criterion of a good interview is short questions followed by long answers, which presumably implies an interviewer who is withdrawn rather than active trying to put the interviewee in the centre. (Alvesson, 2012, p.8)

Localism

This approach emphasizes that interview statements must be seen in their local, situation-specific context. (Alvesson, 2012, p.9)

If the interview is a social encounter, then, logically, it must be analysed the same way as any other social encounter. The products of an interview are the outcome of a socially situated activity where the responses are passed through the role-playing and impression management of both the interviewer and the respondent. (Dingwall, 1997: 56) (Alvesson, 2012, p.9)

Against the neo-positivist, and to a considerable extent also the romantic views on the interview as a technique, localists see it as situated accomplishment (Silverman, 1993: 104). As expressed by Potter (1997: 147) the ‘social structure becomes part of interaction as it is worked up, invoked and reworked’. (Alvesson, 2012, p.9)

Sources of inspiration for localism include ethnomethodology, conversation and discourse analysis. (Alvesson, 2012, p.9)

It encourages a myopic interest in details around what is happening in the interview situation, which is studied as such, and it discourages treating what is being communicated as potentially important information for developing knowledge about broader phenomena, instead seeing it as no more than talk in interviews informed by cultural norms. (Alvesson, 2012, p.10)

interviews are not only used for studying highly personal issues. They are also used in order to get information about what is happening ‘out there’–interviewees will report on social problems, decision making, city life, public sector reforms, organizational functioning, rescue operations and so on, areas where the interviewee is supposed to say something about things not necessarily close to their inner life or emotions. Reports about practices in a neighbourhood, in schools, workplaces, etc. may still be sensitive and include a strong personal element, and they may be difficult to talk accurately and neutrally about in interviews, but is not like reporting about one’s self or other personally sensitive issues in relation to feelings for one’s children or responses to sexual harassment. The critique of emotionalism (and to some extent romanticism more broadly) is not always directly relevant for addressing less personally sensitive issues. (Alvesson, 2012, p.10)

Mixed Positions

The interview then appears, on the whole, to be a valid source of knowledge-production, although the social process and local conditions need to be appreciated and actively managed by the interviewer in order to accomplish valid results. (Alvesson, 2012, p.11)

Perhaps we should not be too keen on being rigorous and coherent here and allow space for double positioning or at least ambivalence. (Alvesson, 2012, p.11)

Riessman (2003), in an overview of interviews as personal narratives, claims for example that ‘narrative analysis takes as its object of investigation the story itself’ (p. 332), which would indicate a rather specific and narrow domain and not the moving outside of it.⁴ But then what is being represented in the narratives ‘tells us a great deal about social and historical processes’, including contemporary beliefs (p. 353) and also ‘individual and collective meanings, as well as the processes by which social life and human relationships are made and changed’ (Laslett, cited in Riessman, 2003: 353). (Alvesson, 2012, p.12)