Research Methods for Information Research
2. Asking questions (and getting research answers)
2.1 Research interviews: forms of interview
Some years ago our project team invited a well-known social work researcher to role-play for us so that we could hone our interview skills before questioning social work managers. At the end of a long afternoon spent in questioning our ‘victims’ and reviewing our performance on video, he admitted:
“I’ve just finished a big round of research interviews and we noticed that there were substantial differences in how much we gathered and what sort of issues came out. I now realise that most of these differences can be attributed to interviewer bias.”
But what can be done to ensure that the results of interviews are as reliable as possible? One factor is the amount of structure deemed appropriate.
In the structured interview, all questions are asked in a predetermined order from a prepared schedule and all the interviewers involved strive to ask each question in the same way and with the same emphasis. A sloppy version of this type of interview is probably the norm in information research.
In the exploratory interview, the question areas are pre-determined but the respondents are allowed some latitude to answer in their own way and the interviewer may probe for more information in promising areas. The exploratory interview is particularly valuable at the early stages of a project to identify issues and find out what concepts and terminology various groups of people use. (For example, ‘information skills’ is a good word to use with education librarians but a no-no for teachers – they prefer ‘study skills’, ‘library skills’ or even ‘problem-solving skills’).
A semi-structured interview is a one-sided conversation in which the respondent is allowed free rein as long as the interviewer considers that what is being said is, or might be, relevant. This works well with experts, who usually have a strong theoretical basis to their thinking which will be explicated by letting them talk. The ‘Catch 22’ here is that although reliability (e.g. reducing interviewer bias and keeping the respondent focussed on the topics of research interest) is likely to be improved by increased structure, too much structure may inhibit responses or produce little of consequence.