The Secret Inductivists

There are a number of research strategies which have origins in the social sciences and relate specifically to social anthropology. Social anthropology is the study of people in their places where they live, their cultural and social forms, and is heavily concerned with method. Usually but not exclusively the social anthropologist goes far away to a different culture of a comparatively primitive people, and the aim of such work is to describe how they live, and what their rituals and culture and social organisation mean to them. The social anthropologist is somewhat likened to a translator from their culture to ours. However, social anthropology can just as well be about places and people nearer to home and indeed at home in the most so called advanced societies. Much of its debate about methodology therefore concerns anyone considering primary research.

One approach in social anthropology is a hypothesis which is grasped by inspired and informed guesswork, and from this researchers get a focus by which to go out into the field. The origin of the hypothesis is less important than carrying it out. Researchers have an idea that so and so is the case, and then they go and get the data to support their hypothesis or otherwise. This is the Popper and Leach method.

A second method that might be used is a hypothesis worked out from a current paradigm. Here the origin of the hypothesis does matter, because it is based on what people currently think is the case. This is associated with Kuhn and paradigm shifts. This is to say that we understand what we do by believing in working and operating hypotheses. These hypotheses must be tested, and we test them by data and falsifiability.

A third method is using a formal model. A good one is to develop ideal types as associated with Weber. This is what I did within my research and I have in fact been to some extent recommending this - suggestions of creating opposites which exist at a theoretical level. In my case I used the well worn ideal types of Church, denomination, sect, but as categories that I wanted to dismiss as inadequate; at the end of my work I came up with a formal scheme that my research did recommend and support - conversionism, traditionalisms, orthodox liberalism, heterodox liberalism.

The problem with having such theory or hypotheses, especially wide ranging and grand hypotheses, and such generally deductive methods, is that you can end up with illustrating a hypothesis rather than critical investigation. This is because it is very easy to find the facts you want to either prove or disprove a general theoretical position. One way to do this is to keep hypotheses practical and well focussed. And we support them with even clearer objectives.

But there is another method that social anthropologists do - and that's the inductive method. This is even when they claim to use some generally deductive method and intend to do such work rigorously.

This is what they do. They go into the field and observe and participate. They have interviews and conversations. They keep diaries; they write. And then they come away with lots of data. And when they get back to the office, so to speak, they spread the data out and look at it. They start to build up a picture of what went on. They see a story of events happening within the data - they may well have realised this in the field (you too may go out to do some work, and see something else happening that seems far more research worthy and interesting). Back at base they then start describing the data in the way it now makes sense. And it is from the data that they see theoretical points. They see the hypotheses and the theories afterwards!

I admit that this happened to me. I actually observed and then changed my own research. The ideal types method seemed best for my purposes but only later on. It was then that I organised the issues and the hypotheses and the rest to fit the kind of data I was collecting.

We can call this inductive method "model building" - you build up from the bottom. When you research plan you do some research to start with. You have to know the answers a little to set out the questions. So why do you have to do some research in order to plan it? Because the inductive method always comes to us, it always undermines the best laid plans. In other words you go out to the field to find the questions. And this is what so many social anthropologists do.

And then you write the work as if you started with the hypothesis and ended up with data to prove or disprove (crudely speaking) the hypothesis or hypotheses used. We all know that this is a bit of a con because it was only from something to do with gathering data that we actually came up with the hypothesis.

So writing research is a bit of fiction. It isn't really a history of what happened, it is a presentation of a research method - but the method you illustrate is almost certainly not the full story.

Just as a social anthropologist is compared with a writer, if not quite a writer of fiction, then so your task and my task is to write well, to represent what we see and turn it into a good story.

Using Kuper, A. (1983), Anthropology and Anthropologists: The Modern British School, London: Routledge and Kegan Paul, 202-204.