3.1 Interviews

Interviews can be quite structured, question-and-answer interactions. This kind of interview is similar to a survey or questionnaire that is being asked by an interviewer rather than filled in by participants on their own. There is little room for any follow-up questions as the interview protocol usually needs to be followed quite precisely by a number of different interviewers. An example might be a community needs assessment where several interviewers are sent out into a community to gather information about people's needs and aspirations.

At the other end of the spectrum interviews can also be quite unstructured and closer to a conversation. The research question might be the topic of conversation, with the participant sharing what they think is relevant and the interviewer drawing them out by listening and asking questions that explore parts of the conversation in more depth or ask about things that the participant may not have talked about.

In-between these structured and unstructured interviewing methods, are semi-structured approaches where the interview revolves around a series of key questions or interview topics, but still puts the participant in the 'driver's seat' about what they talk about in response to these questions. The interviewer's role is again one of listening attentively and asking follow-up and clarification questions.

Interview questions

When you're developing questions for a structured or semi-structure interview, begin by asking quite general questions first; for example, what do you like about living in this community? The following, and any follow-up, questions can get more specific. When we interviewed Māori about their experiences of health care we had five topics that were covered in semi-structured interviews:

• • What is Māori health?

  • Differences between Māori health and the health of the rest of the nation

  • Personal experiences with general practitioners and other healthcare providers

  • Experiences of family and friends

  • Traditional Māori health practices

The first question or topic allows you to scan your participants' environment so that you get an idea of who they are and how they think about life. This first question might also be quite an affirming question (i.e., positive rather than negative questions). These type of questions open up a whole vista that participants can talk about, including what's good, not so good, and plain bad. If you begin with a general, negative question you may signal to a participant that you only want to hear about the bad things and they may chose not to talk about their positive experiences or views (whereas participants usually don't hesitate to talk about the good and bad in response to an affirming question).

Number of people at one interview

Interviews can be between the researcher and one or two participants, or the researcher with a group of participants. Usually when there are many participants involved the interview will be less semi-structured so that the participants talk and share with one another, with the researcher as the facilitator of this process. A focus group interview will involve around 6-8 participants, whereas a hui (meeting) can involve many more participants. In a large group participants may have small group discussion that they then share with the larger group, or they may stay in the larger group so that everyone can hear what's said.

Recording and transcribing

I recommend that you get participants' permission to record their interviews. If participants don't want to be recorded, you have two options: take extensive notes of the interview (although they may not want you to do this either), or don't interview them. (Another option is to commit the interview to memory and this is a preferred method in some Indigenous cultures where orality is highly valued.) What you decide to do is up to you but it's hard to continue on to an analysis or reporting of interview findings without some recorded data about what was said. Recordings of interviews should be transcribed word-for-word. We often return transcripts to participants who've been individually interviewed so that they can check and add or delete information. People often don't like transcripts because they worry about the way they talk (e.g., that it doesn't make sense), so if you plan to return transcripts you need to prepare your participants for what talk normally looks like (e.g., convoluted, changing topics mid-sentence) and put them at ease. You should not return transcripts from group interviews as it's inappropriate for people to receive a record of what others in their interview group said.

Total number of participants

If you're interviewing individuals or 1-2 people at a time, then the usual guide is to interview 20-25 people or to keep doing interviews until you're not really hearing any new information. If you're interviewing people in focus groups then around 4-6 focus groups may provide you with enough information for your first research project. If you are convening hui then you may want to give people the opportunity to come to one of three hui that you call.

Facilitating rather than interviewing

The researcher can step back more into the background in larger group interactions by using tools that encourage participants to engage more directly with one another. For example, memorywork involves the members of a focus group taking turns to share a story about a particular topic, drawing out themes from the stories told, and then discussing these themes. Vera Keefe-Ormsby used memorywork to talk with Māori men about the closure of the Whakatu Freezing Works (abattoir). On their first night together they told stories and discussed working at the Works, on their second night they told stories and discussed the Work's closure, and on the third and final night they discussed the impact of the closure ten years on.

Ross Conner uses a technique called 'Draw the Path' where each participant in a group gets a marker pen and the group is asked to describe their journey with an initiative (e.g., service, or community activity) by writing and drawing on a large sheet of paper. A line is drawn on the paper (by the researcher or the group itself) that has 'beginning' at one end and 'now' at the other. The group then talk and draw and fill in their timeline while the researcher stands back. Once they have finished the group present back what they have drawn and the researcher then gets the chance to ask clarification questions. We used this method in the development of an evaluation framework for Te Puni Kōkiri's Whanau Social Assistance Programme. Click on the image below for a Draw the Path illustration of a gardening growing cycle, drawn by young people at Te Mahurehure Marae, Auckland.

Analysing interviews

There are many ways you can approach the analysis of the information you collect from conducting interviews. For example, analysis can involve the coding of participant responses and counting the frequency of different themes that emerge, or it may be more about allowing common themes to emerge through the reading and re-reading of interview transcripts by the researcher. My advice is that you find a way of analysing your interviews that feels right to you. This might be Interpretive Phenomenological Analysis, or content analysis, or discourse analysis. There are many schools of thought and it may be that you want to try one that looks interesting to see if it works in a way that you and your participants think is appropriate.

The analysis you choose will probably be tied to a methodology for conducting your interviews, so you will need to do a background check to ensure that you're following correct procedure from designing, to undertaking, to analysing your interviews. What I've given you above is just a broad outline of how interviews might look.

Many Indigenous researchers talk about the importance of dreams and other signs as informing their analysis, and also their research project. These are usually not talked about in Western research texts but are nonetheless culturally valid additions to your analytical toolkit.

Thematic analysis

If you are going to do interviews, my recommendation is that you begin with a semi-structured interview guide (5-6 questions or interview topics), interview 15-20 people individually or in small groups (2-3 people), and do a thematic analysis on their responses, as this method of analysis is not tied to any particular qualitative theory. Being familiar with your interview data is a big part of any qualitative analysis so you should carry out the interviews yourself, as well as transcribe them yourself, then read and re-read the transcripts until you feel you have a good idea of what sort of themes are emerging from what participants have said. Your next task is to code the interview transcripts for these themes. There are software packages that you can use to do this but when you're starting out I suggest you go about coding with different coloured highlighter pens on a hard copy of your interview transcripts. This is a process of trial-and-error and, in my experience, it's usually not until my third attempt at coding transcripts that I feel like I've gotten it right. And this is usually after the first and second coding attempts have each been written up but I'm still not been happy with how they've turned out. So don't hesitate to go back and start again if you have a fresh insight. And remember to leave yourself lots of time to do this in your research plan. This can't be rushed.

When you have themes that you've happy with (and it may just be 6-8 main themes), then write about each theme and use portions from the transcripts to illustrate what you're writing about. Within a theme there might be participants who agree or disagree (and sometimes participants will contradict themselves in different parts of their interview), so explore this as well. You can also say how many participants talked about each theme. See the article on research ethics by Flicker and Worthington (2012) for an example of thematic analysis of semi-structure interviews.

Analytical hesitation

There are two hesitations I've heard from people about a thematic-type of analysis.

The first is that it cuts people's kōrero (talk) up into disjointed portions. If this is a concern for you then you may want to interview people using more structured questions, and do a thematic analysis for each question. This maintains the integrity of participants' responses to particular questions. If you want to do semi-structured or unstructured interviews and look at participants' kōrero as a whole then you may want to chose a narrative or ethnographic method such as storytelling.

The second hesitation is about analysing people's talk when researchers see kōrero as something too sacred, or tapu, to be subjected to analysis. When this happens the temptation is just to present what people say for readers to make meaning of. Unfortunately this is a lot of work for even the most well-intentioned reader, and those with less good intentions may make meaning that the participants would find objectionable. There is some middle-ground whereby you can present what people say but then wrap around that your own analytical lens to guide readers about how to understand and interpret participants' talk. In this way you are offering your interpretation, not THE interpretation, and you are also offering your participants the opportunity to see what you have made of what they have said in a way that isn't a re-working of their own words.

References

Helpful links

Designing and Conducting Semi-Structured Interviews for Research. Prepared by Ted Zorn, Waikato Management School.

Robert Wood Johnson Foundation. Qualitative Research Guidelines Project. Interviewing.