Cousin, G. (2008) Semi-Structured Interviews. In: Cousin, G. Researching Learning in Higher Education: An Introduction to Contemporary Methods and Approaches, pp.71-93. London: Routledge.

Semi-structured interviews allow researchers to develop in-depth accounts of experiences and perceptions with individuals. By collecting and transcribing interview talk, the researcher can produce rich empirical data about the lives and perspectives of individuals. (Cousin, 2008, p.71)

Unstructured interviews are where the researcher guides naturally occurring conversations. (Cousin, 2008, p.71)

Unlike the structured interview, the interviewer is expected to adapt, modify and add to the prepared questions if the flow of the interview talk suggests it. (Cousin, 2008, p.72)

Semi-structured and unstructured interviews attempt to grapple with complex experiences. (Cousin, 2008, p.72)

Some research designs will include both the equivalent of “counting the dollar income” to get a broad, contextual picture and in depth interviews to drill down into how people operate within this context. (Cousin, 2008, p.72)

Fontana and Frey (2000: 645) point out that “asking questions and getting answers is a much harder task than it may seem at first. The spoken or written word has always a residue of ambiguity, no matter how carefully we word the questions and how carefully we code and report the answers.” (Cousin, 2008, p.73)

interactional event in which meaning-making is in situ and the product of both players in the interview rather than that of the skilful transcript analyst after the event. The interview, write Alldred and Gilles (2002: 146) “is the joint production of an account by interviewer and interviewee” and again, to quote, Holstein and Gubrium (1997: 14) “meaning is not merely elicited by apt questioning nor simply transported through respondent replies; it is actively and communicatively assembled in the interview encounter.” (Cousin, 2008, p.73)

unstructured or semi-structured interviews are best conceptualized as a “third space” where interviewer and interviewee work together to develop understandings. (Cousin, 2008, p.73)

Holstein and Gubrium (1997) argue for there to be dual attention to both the how and the what of an interview. This requires an interest in both the content of the interview and in the ways in which such content is assembled. (Cousin, 2008, p.74)

The active interviewer sets the general parameters for responses, constraining as well as provoking answers that are germane to the researcher’s interest. He or she does not tell the respondents what to say, but offers them pertinent ways of conceptualizing issues and making connections— that is, suggests possible horizons of meaning and narrative linkages that coalesce into the emerging responses. Holstein and Gubrium (1997: 125) (Cousin, 2008, pp.74-75)

The researcher must listen out for how the “circumstances of the interview” (Atkinson and Hammersley, 1994: 113) might be shaping what the interviewee is saying. In short, are there consequential power asymmetries between interviewer and interviewee? (Cousin, 2008, p.75)

Broadly, there are three ways of addressing researcher positionality. Firstly, where it seems a good move, efforts can be made to establish some form of commonality between the interviewer and the interviewee. (Cousin, 2008, p.75)

Secondly, the interviewer needs to do his/her best to minimize the power present in the interview by, for instance, disclosing their own relevant experiences and by facilitating an exploratory thrust rather than an information prospecting one. (Cousin, 2008, p.76)

Thirdly, the researcher needs to take a reflexive stance by problematizing positionality throughout the interview process. More generally, it is important that the interviewer develops a conversational style, building a trust relationship with the interviewee. (Cousin, 2008, p.76)

For most interviewers, the quality of the conversation between the interviewer and the interviewee is in large measure dependent on the rapport building capacity of the researcher. (Cousin, 2008, p.76)

Do not enter the interview stage with a deficit model explanation of your interviewees or a political investment in over-determining them as different and “other.” We are always more than the categories assigned to us and an interview should not be in the service of a narrow hunt for social difference data. The point of an interview’s indepthness is to explore subtleties, nuances, uniqueness and singularity alongside possible generalities and commonalities across groups. (Cousin, 2008, p.77)

Non-linguistic communication will impact on the interview: this includes interpersonal space, use of pauses and silences, tone of voice, body language, dress and other modes of presentation of self from the interviewer (Fontana and Frey, 2000: 660–661). (Cousin, 2008, p.77)

Their perspective accepts that interviews make meanings in situ but it also defends researcher authority in the shaping of the interview and the analyzing of the data to understand something of the social world. (Cousin, 2008, p.78)

An interview “requires that the interviewee consents to the interviewer to ask him or her questions on an agreed topic and to use the resulting transcript for research purposes.” This blunt delineation of the interview from Denscombe (2007: 173–174) (Cousin, 2008, p.78)

Informed consent often needs to be ongoing rather than a preliminary stage of the interviewing; this includes letting the interviewee see the transcript and comment on your analysis if you can. (Cousin, 2008, p.78)

• Have you avoided smoothing your interpretation? For instance, have you been sensitive to variation of experience and viewpoint within groups as well as across them?
• What has determined your decision about numbers of interviews? Can you defend this?
• Have you shared your emerging ideas with the interviewees throughout the interview?
• Did you build rapport with the interviewee?
• Have you asked a colleague to check some of your transcripts/tapes to explore rival explanations and interpretations?
• Have you kept a diary to capture your reflections, theoretical leads and so forth?
• Does your research report avoid cherry picking quotes?
• Have you displayed your own questions as well as responses to show a developmental dialogue?
• Have you explored alternative data sources to strengthen your interpretations? (Cousin, 2008, p.79)

“purposive sampling.” This simply means recruiting people on the basis of a shared characteristic which will help you in your inquiry. (Cousin, 2008, p.79)

Theoretical sampling involves two moves. Firstly, you must analyze your data as you interview (good advice for any qualitative research) and second, you begin to make judgments on the basis of the emerging analysis about how many more people you want to interview and what kind of experiences you would like them to reflect. Theoretical sampling means that you talk to more people to go more deeply into the issue, not to accomplish coverage of all possible sources of variation. (Cousin, 2008, p.80)

Let the interviewee know that you are hoping for an event that will not be overly formal. Explain that you want to interview them because they have the expertise and experience you are seeking to learn from. (Cousin, 2008, p.80)

I usually give the interviewee brief typed information about the purpose of the research project, the confidentiality of the interview and my contact details. I secure consent for taping the event and for using the eventual transcript (duly anonymized), letting the interviewee know that I will send it to him or her to check and to add any further comments (much of this will need to be repeated at the beginning of the interview). (Cousin, 2008, p.81)

This acknowledges that people do not always have a consciously worked out reason for why they do certain things. If the interview aims to be developmental, you need to come up with questions that are more than information seeking. (Cousin, 2008, p.82)

You need to prepare for an interview by reading as much as you can about the issue so that when you hear or offer provisional ideas, connections and promising conceptual hooks, your sense-making is from a scholarly basis. This reading will also help you to formulate the direction of your questions for the interview schedule. (Cousin, 2008, p.83)

The semi-structured interview schedule is always a working document. In the course of interviews, you might add to or amend questions as interviewees alert you to the need to do so. (Cousin, 2008, p.83)

Perhaps it is best to think of the interview as an interactive performance that needs some preparatory scripting and stage management. The interviewer has to acquire a repertoire of questions (prepared and impromptu) for encouraging a good, exploratory conversation. Technique needs to be accompanied by a scholarly understanding of the field so that the formal questions and the ad libbing are not in a theory-free zone. (Cousin, 2008, p.84)

A semi-structured interview has between five to eight main questions. (Cousin, 2008, p.84)

Tour questions are common openers because they tend to be unthreatening and a fruitful way of getting interviewees to talk about their experiences. (Cousin, 2008, p.85)

A further fruitful main question asks the interviewee to identify experiences or events which stand out in some way. Like Rubin and Rubin (2005: 161), I find this kind of question often yields very rich and unexpected data. Similarly, eliciting “Iconic moments” (Rubin and Rubin, 2005: 132) can produce vivid accounts. (Cousin, 2008, p.85)

Further suggestions from Rubin and Rubin (2005: 161) include posing hypothetical questions. These can take the form of an imagined future (Cousin, 2008, p.85)

Asking interviewees to compare and contrast two sets of experiences prompts them to think about what is distinctive about each (Cousin, 2008, p.86)

The purpose of probes is to “help you manage the conversation by regulating the length of answers and degree of detail, clarifying unclear sentences or phrases, filling in missing steps, and keeping the conversation on topic” (Rubin and Rubin, 2005: 164). (Cousin, 2008, p.86)

• That’s very interesting. Can you tell me more?
• That must have been difficult. Can you tell me more?
• Do you have an example of when that happened?
• Did that happen often? How many times?
• Could you clarify? Perhaps with an example?
• What does “x” mean?
• Then what happened?
• How do you see that as related to the topic?
• We can come back to that later if you prefer (Cousin, 2008, pp.86-87)

Other probes can be non-verbal such as nods, “uh ha’s” and encouraging silences to signal that you want to hear more. (Cousin, 2008, p.87)

The success of the interview centers on the interviewer’s ability to follow hunches, hints and openings through good follow-up questions. (Cousin, 2008, p.87)

As the interview time runs out, plan to end punctually. If you think that the interviewee has more to say, you can arrange to meet again or to have a follow-up telephone conversation or email exchange. Ask the interviewee if he or she has any questions. (Cousin, 2008, p.90)

Finally, you thank the interviewee and repeat that he or she will get a transcript; specify a deadline by which you would like comments back, failing which, you will assume consent to quote from it. (Cousin, 2008, p.90)

Once the interview is over, you write your own field notes about the event to add to your data. This will include reflections on the questions and answers, gaps you think you didn’t fill, any feelings you had about the setting and interviewee and so forth. (Cousin, 2008, p.90)

Most transcripts are verbatim and include pauses, repetitions and idiomatic expressions (such as “innit” or “you know”). Walford (2001) argues against complete transcription, preferring instead to listen to the recordings and to take notes from them. This enables Walford to have a stronger sense of the interview event and to avoid treating people’s testimonies as their last and definitive word on the topic. (Cousin, 2008, p.90)

In displaying some of the transcript in your report, use the quotes convincingly: they must not look laundered or cherry picked. Incorporate your comments to show the developmental dynamic of the event. (Cousin, 2008, p.90)

Holstein and Gubrium (1997: 127) explain: ‘The goal is to show how interview responses are produced in the interaction between interviewer and respondent, without losing sight of the meanings produced or the circumstances that condition the meaning-making process. The analytic objective is not merely to describe the situated production of talk, but to show how what is being said relates to the experiences and lives being studied. (Cousin, 2008, p.90)