Making the connection

Guest post by Kathryn Oliver

About 2 years ago, I had one of those ‘Eureka’ moments that totally changed my life. Genuinely. It was right up there with finding out about Oyster cards, or washing machines, or something.

At the time, I was a PhD student in my first year, working on a fairly standard project about developing health indicators. As a project, it was fine – about the use of evidence by policy makers, one of my main interests, and I was getting lots of experience in survey design. But for years, I’d been kicking round ideas in my head about the importance of personal relations. Didn’t they really explain nearly all human behaviour? Weren’t peer effects important for the spread of obesity or smoking? Wasn’t social capital important for mental health?

I’d been living on my own in London for a year or two and had found myself pondering the role of human relationships more and more. Of course, I had friends and relations, but I also liked being known by the man in the newsagents and the end of the road, and saying 'hi' to the neighbours. Did they count, I wondered? Would these relationships be enough to protect me from isolation, or going ballistic on the tube?

Imagine my delight when, attending a Social Network Analysis seminar day, run by the Mitchell Centre at the University of Manchester, I discovered an entire body of research – methods, philosophy, approaches – which looked at connections between individuals using formal statistical methods. Finding out that other people had had similar ideas to me, and had developed dedicated research methods to investigating these ideas was probably one of the best research moments I’ve ever had.

Unlike traditional statistics, network analysis does not treat individuals (whether bridges, policy makers, or swingers) as independent. Instead, any ties between actors are identified, described quantitatively and/or qualitatively and mapped. The statistics used are based on graph theory, but you don’t have to understand it to admire the elegance and usefulness of network analysis. Depending on the relationship collected, people’s attitudes, behaviours, health outcomes and more can be predicted.

For me, this is really the missing element from a lot of public health research. It can be used to identify good targets for research, or opinion leaders in secondary schools, so more targeted messages can be produced and sent out. It allows us to understand, describe, and analyse the social context within which individuals live. And, of course, make beautiful pictures. 

Example of Social Network Analysis diagram.

People have used network analysis to study all kinds of things – it’s very popular in the business world to identify ‘future leaders’ or ‘people who make things happen within my business’. Researchers have compared US senators voting patterns to cows who lick one another.

My PhD changed quite a lot after this seminar. I ended up using a combination of social network analysis and ethnography to study where public health policy makers found evidence, who the main sources of evidence were and how evidence was incorporated in the policy process. For years, academics in my field have been talking about the importance of interpersonal knowledge translation and how policy makers prefer to get their info from real people. Now I’ve been able to add my own tiny part of the story, come up with new research ideas on the basis of my findings, and learn a niche method (always useful).

My boyfriend still calls them snog webs though.