Thursday, 29 August 2013

Conducting qualitative interviews

I am working on a publication on qualitative interviews – but it’ll be in German and it’ll take ages to complete because it is one of those spare-time not-for-profit projects. Since a friend is about to embark on a round of interviews we have designed together, I have decided to summarise some bits in English. Here they are! 


A qualitative interview is not a questionnaire that would be administered mechanically. It is a more or less structured conversation on issues that are important for your research. To obtain the data on the interviewee’s experience and opinions you are looking for, you must meet her or him in an open, attentive way. (For the sake of easier reading, let’s say the interviewee is female.) You need to inform her about the purpose of the conversation at the outset, try to create a comfortable atmosphere, and listen carefully and respectfully.

Do remember, though, that this is not an everyday conversation– it is an interview that serves a specific purpose: you want to obtain data for your research. That is why you need an interview guide - i.e. a document that reminds you of what you need to say and ask. You should talk only as much as you need to make the interviewee feel comfortable and ask your questions. Also, keep your reactions under control so that they encourage the interviewee to tell you more – even if you disagree with her.

Practise active listening: Keep eye contact with the interviewee and show interest in what she says. When she says something that seems particularly important, paraphrase or summarise her statement to make sure you have understood. For instance: “So, I gather that the leaflets distributed in the training course are difficult to understand, and that is why you have not used them after the training. Is that correct?” When she says something that is vague or difficult to interpret, ask for concrete examples: “You just said the project was not sustainable. What exactly do you mean by ‘sustainable’? Can you give some examples?” 

Keep track of time. Try to ask all the questions on your interview guide. In my experience, it is wise to limit the guide for a 1-hour interview to some 10 questions – you will ask probing questions in-between. It is OK to interrupt the interviewee – especially if you tell her at the beginning of the interview that you might interrupt her. Do make sure you have at least 5 minutes at the end of the interview for an open question: “What else would you like to add?”
Difficult cases: Some people are very talkative. If they say interesting and useful things, don’t rush to interrupt them. But keep an eye on the questions on your guide to make sure they answer all of them, or at least those you have prioritised beforehand. If they drift too far away from the issues that are interesting for your research subject, stop them – in a friendly but firm way. Keep smiling (even if you're talking into a headset)!

I have also come across people who are very good at delivering heaps of content-free development jargon – do not give up! Just keep asking them for concrete examples. This is also a good recipe when you deal with laconic interviewees, common in Germany, a rather sober nation (I am allowed to say that because I am German). If you rephrase your questions, show enormous interest and ask for examples, you will make us talk, eventually. I have encountered a different set of difficulties in Myanmar, where no-one seemed willing to come forth with any kind of statement that would imply the slightest sense of negativity. Rephrasing questions and probing helped a little bit – but what really made a difference was to enrol a local researcher to conduct the interviews. Remember that when you conduct interviews far away from home.

Finally, a note on risk reduction. A good interview is likely to stimulate the interviewee’s thinking and stir her feelings. Some feelings can cause harm, such as re-traumatisation which occurs when a conversation conjures traumatic past experiences. This can happen in research on violence against women. The time-tested WHO/ PATH manual (click on the link to find it) offers guidance as to how to limit the risks for everyone involved.

No comments: