Research methods – why, how, why?

I presented my first go at an Action Plan to the group last week and although the clarity of the message on my slides possibly suffered from being put together pretty late last night, I felt by the end of the feedback and discussion with the rest of the group I had quite a good idea of what I am aiming at. After some persistent challenging from Sheila I realised that my plan to get more familiar with the subject of my research before narrowing in on a more focussed research question, which seemed like common sense when I wrote it down, was actually the product of exactly the kind of thinking that I have been trying to avoid. It is the antithesis of design thinking, of Agile development, of rapid prototyping, of any kind of iterative process. It’s waiting and waiting until the very end to produce something and then expecting that thing to be perfect. It assumes we’re dealing with complicated, but not complex systems (see Nesta’s report on Systems Innovation). Basically, it’s old school.

It was when I realised the importance of having not just a research question but also a hypothesis that everything fell into place. Like with prototyping, it doesn’t matter if the hypothesis is not right – in fact it’s probably better. Discovering that something is wrong gives you the opportunity to ask why and to learn more about what you’re trying to develop. This revelation then helped me to make sense of the research methods described in Real World Research by Colin Robson. It gives a basic outline of different research methods that can be used to study real world situations, including interviewing, observation and surveys. For each methods it describes subcategories as well as when these are appropriate, what activities they comprise, options for conducting them and advantages and disadvantages associated.

I’ve been really interested in these kinds of research techniques for some time, but never had an opportunity to use them. All the time I’ve wondered how it is that you can use observation and interviewing, which can come in various forms ranging widely in how structured they are, to produce evidence that is any more than a systematic kind of social value judgment. Once you have asked your questions and made all your notes on the various behaviours, then what? You can’t just take each sentence or action, or even pattern of actions, in turn and analyse all their various possible meanings in the hope of getting closer to some kind of insight or truth. But when I read Robson’s process of analytic induction, things started to fall into place:

This is yet another type of iterative development process! Pick a hypothesis, any hypothesis, and in doing so give your data something to fight against. Ask of that hypothesis, what does true look like?  If the data looks different then keep refining and gathering more data until you the truth of the data matches the truth of the hypothesis.