Friday, November 18, 2005

The Curse of the Semi-Structured Interview

Quite a lot of my time recently has been taken up by reviewing. Most of the activity is reviewing research grant applications for one of the research councils in the UK, but other research bodies also ask me for reviews. I like reviewing. It's good to see what is being done out there in the academic world; it gives me ideas; and most importantly it reminds me of the mistakes we most commonly make when writing research proposals.

The key components of a good research proposal are a good research question, one that is going to drive the research along, and a set of methods that are appropriate and well-specified. Since I teach research methods to law graduate students, I am constantly repeating these points. (I may have reached the stage where I dream about them.) I know that students won't always fully appreciate this the first time round, but they will get it eventually.

It's the full-time academics who don't get it that surprise me. I've mentioned two components: research questions and methods. They are the basis for a good and successful proposal. Research questions fail because they are ambiguous or too vague or, in some cases, aren't really questions. Methods, well, methods. In the socio-legal community there is generally an appreciation of the role of methods in research. How exactly are you going to get your information? Who from? And so on. Too often one sees "the data will be collected using semi-structured interviews". And that's it: nothing else. This is not good. I sometimes believe we live under the curse of the semi-structured interview. It has become the get-out clause for many researchers. One doesn't have to specify too much and basically one will think of something at the time of the interview. That's the kind of impression it creates.

Of course there is a place for this type of interview. But researchers must realise that they need to supply more information about the sorts of things they will ask about, to whom, etc. And also provide some idea of how they will analyse their data. Without this the reviewer is left in the dark.

I recently visited a research council and watched over 30 research grant applications being appraised. The clear losers failed for the reasons I've outlined. It wasn't the novices or junior applicants that were the worst cases, rather it was senior people who gave the impression of not bothering too much. When you are asking for money, whether it's £1500 or £400,000, you better bother a lot.


No comments: