Research Methods for Information Research

2. Asking questions (and getting research answers)

2.2 Research interviews: the five main skills

Once you have chosen the type of interview, try to get reliability of performance by systematically applying our preferred definition of the research interview: ‘a conversation conducted to obtain research-relevant data.’ Anything that helps to do this is appropriate; anything else should be discouraged or limited.

The interviewer can do much to create good interview conditions, ‘train’ the respondent to perform their task and reduce incomprehension or resistance. In the end though, we are largely reliant on how meticulous the researcher is in the five main areas of research interviewing – asking the questions, listening to the replies, ensuring that the respondent answers in full, recording the answers, and (usually later) reporting back the overall findings. All of these areas are problematic, but most people can get to perform better in each one through focused training and reflection on their own performance.

I was introduced to a rule driven approach originally contrived by Michael Brenner to prepare market researchers for their task. What follows is a version of the Brenner process4, shorn of a few of his rules. The essence of this approach is that the interviewer should be proficient in five main areas. The interviewer should:

  • ask the question – read the question at ‘slow normal’ speed and in a neutral tone, but with emphasis on certain words to help the respondent understand the question.
  • listen to the reply – this is usually the easy bit because the replies should be interesting if your questions are good. Don’t get distracted if the respondent sounds boring – the most monotonous replies can turn out to be full and revealing.

These two steps are fairly straightforward, but if video evidence is anything to go by, are more honoured in the breach than in the performance.

Life gets interesting in the next stage, using non-directive probes and prompts. You should:

  • Ensure that the respondent answers in full by using non-directive probes or prompts. The aim here is to get answers to the question without influencing the answers accidentally by putting words into people’s mouths or directing them towards one kind of answer. If you want someone to answer a question in more depth, try a non-directive probe, such as:
    • the silent probe – you ask a question and get a superficial or cryptic answer. Wait, with pen poised. This is all that is usually needed to get the respondent to elaborate. Even a ten-second pause can feel like eternity if you are on the receiving end.
    • the encouragement probe (semi-verbal) – you make a noise to show that you are listening and interested – such as ‘ahah’ or ‘erhem’. This is a concession to the one-sided nature of the conversation. It can be disconcerting if you are being interviewed and are talking away but getting no response at all. The sorts of grunts advocated here seem ludicrous when written down but work well in practice. As soon as you move away from non-verbal cues you are in danger of influencing people’s answers. For example, if you say ‘yes’ from time to time, does the respondent think you are agreeing with the point of view expressed?
    • the elaboration probe – you can ask a respondent to explain an answer, but take care not to offer your own interpretation in case they just take the easy way out and agree to your version. Often, all you have to do to get someone to be clearer is to repeat the part of the answer that puzzles you. If that doesn’t work, add ‘What did you mean by that please?’
  • You ask a question and the respondent doesn’t answer. Try one of the following non-directive prompts:
    • give time for the person to answer. Strictly speaking, this is not a prompt, but remember that your respondents are hearing your question for the first time. They may need time to think; keep silent to encourage them to do so.
    • dealing with a challenge – if the respondent asks whether your question means A or B you have a quick decision to make. Your main options are:
      if you think that the respondent hasn’t fully grasped the question, you should repeat it
      if the respondent asks which meaning is intended, try saying ‘Whatever it means to you’. This usually works (but make a note of the problem – you may have to revise the question later)
      if the respondent is still struggling, reword the question (and
      make a note of your exact words – so that you can review the question and answer afterwards, and if necessary discard both)
    • predetermined prompts - pilot the interview questions properly in advance to make sure you are aware of potential hazards. You are then in a position to prepare definitions or clarifications in advance and give them when asked.
    • challenge prompts – people aren’t always consistent in their replies during an interview. If you think that a reply is inconsistent with something said earlier, refer back to the previous answer and remind the respondent what he or she said. This is usually enough to get an explanation.

Now we are ready for what should be the easy part, recording the answers – unfortunately the huge beginner’s mistake is to forget to do this (we have huge amounts of videotape evidence to prove it!) You should take notes even if you are recording the interview – this will help you make sense of the replies at the time and give you a quick route into the interview afterwards. Remember that it takes about five hours to fully transcribe one hour of a recording. If you are not recording by other means, try to capture as much as possible of what people say in their own words. This should be the hard job. If you are finding it easy you are probably not probing enough – so why bother interviewing?

Finally, the interviewer should feed back complex replies to make sure that you have understood them and to prompt further comment. This takes confidence but is well worth doing.

This approach was offered as a counterbalance to the increasingly popular empathetic approach, where the researcher aims to engage with the interview respondent in constructing a story together.

Empathetic interviewing can be a powerful way of exploring issues. It can also readily become an explicitly political type of engagement. Fontana and Frey (two leading advocates of the empathetic approach) assert that the interviewer then “becomes an advocate and partner in the study, hoping to be able to use the results to advocate social policies and ameliorate the conditions of the interviewee,” before tellingly adding that “The preference is to study oppressed and under-developed groups.”5

Positivist? Who me?

Several thoughts are sparked by this description: the approach might be (and often is) equally readily applied in many consultancy settings, although this might come as a surprise to the more politically radical social scientists. The information consultant hired in to help sort out future information services is hardly the detached neutral enquirer. Secondly, how can advocates of this approach guard themselves against slipping into a paternalistic view of ‘studying’ oppressed and under-developed groups or of plunging into self-indulgence masquerading as research? Thirdly, why do people who espouse the Holy Grail of advocacy and partnership find it necessary to belittle the self-awareness of ‘more traditional sociologists and researchers’. To quote Fontana and Frey again, the traditionalists apparently follow a ‘how to’ approach to interviewing “where the illusion exists that the better they execute the various steps, the better they will apprehend the reality that they assume is out there, ready to be plucked.”6

I have yet to come across a researcher who holds to this illusion or who regards the interview interaction as anything less than a complex, messy and continuously problematic mutual discourse. However, some researchers probably do consider that an approach to consistency in questioning may be of use in ensuring that significant conversational heads or question areas (as seen by the interviewer) are explored as an alternative to wallowing in a self-created (or even mutually-created) conversational mire. This may be self-delusional, but is it likely to be any more so than imagining that the sociologist or researcher can expect to enter into the world of their interviewees to the point where they can ameliorate the conditions of their fleeting ‘partners’?

Two important dimensions of research interviewing are worth stressing here:

  • the importance of empathy and the difficulties inherent in establishing empathetic engagement within the confines of an interview, whatever the setting;
  • the need to seek a balance between breadth and depth, information and understanding, empathy and ‘perspective’, engagement and detachment throughout the interview process. This requires the interviewer to make careful decisions when constructing the interview schedule (if this is appropriate), crucially to make on the spot choices at various points during the interview when the respondent raises issues by asking questions or giving information, and yet again when the interviewer becomes interpreter in analysing the interview results.

4. BRENNER, M., BROWN, J. and CANTER, D.V. (eds) The research interview: uses and approaches New York, USA: Academic Press 1985 ISBN 0121 315800

5. FONTANA, A. and FREY, J.H. ‘The interview: from neutral stance to political involvement’ in DENZIN, N.K. and LINCOLN, Y.S. The Sage handbook of qualitative research (Third ed.) London: Sage 2005, p. 696.

6. Op.cit. p.707.