Saturday, 1 March 2008

A mention in the Economist

I just got an email from our university's Press Office to say my work on how drivers' overtaking behaviour gets more dangerous when a cyclist wears a helmet was mentioned in a recent article on risk in The Economist. It's interesting how that project of mine has been viewed and used by different people. One of the main findings was that a behaviour intended to reduce risk (putting on a bicycle helmet) might paradoxically increase one's overall level of risk because drivers react to its presence by changing their behaviour. I've seen this finding used in many discussions -- some people find it a curious datum, others feel it backs up their own experiences, and some people loath my findings, usually because they are starting with the 'common sense' position that bicycle helmets must be a good thing.

But I've also seen that research used in broader contexts. Indeed I've seen it used in relatively extreme right-wing libertarian writing to justify an argument for having no state intervention in people's day-to-day behaviour. It just goes to show that when you do research and put your findings out into the world, you can never be sure exactly what's going to happen to them.

But for all that, I have to say that it's very satisfying that work of mine is being used by people. It's just a shame that any use of one's research by people who are not themselves professional researchers "doesn't count" under our government's Research Assessment Exercise. Academics writing in endless circles about one another's work is "good research"; a grand theory which excites everybody but then proves to be completely bogus after two years is "good research", as there will be lots of papers supporting it, then a second raft of papers slamming it, giving the high number of citations that research assessment focuses on to a large degree. But studies which excite the public, lead to media discussion or even change public behaviour for the better don't count as having had any impact at all -- our research is only "good" if other academics write about it. But we'll moan about that another day. It's far too sunny a Spring day for caviling now. I'm off to walk the dog instead.


incredulous said...

I've come to this thanks to a referral by the omnivorous Tom Vanderbilt. But, I'm writing because I'd seen a year ago on a cylcing listserv a reanalysis of your study, discussed by Vanderbilt in "Traffic". Odd yow you've kept at arm's length the substance in your appreciation for your research being noticed!! On the chance that someone will follow here, I'll report that at issue is whether a statisticaly significant increase in overtaking distance by drivers you interprete to be wary of the cyclist's stability is substantively significant also. To wit: Does 15cm lesser passing distance by a helmeted cyclist on an average of 150cm really make the helmeted bicyclist less safe? (Exact #s only vaguely remembered.)
But, entertaining stuff, and I'm enjoying the entries on your blog 30 years after working in highway safety myself, and just a few after returning to somewhat serious cycling, though less than your Dad's.

Ian Walker said...

Thanks for the comments. I strongly suspect a shift in distance is meaningful. Here are a couple of facts we can't argue with: (1) millions of times every day, drivers pass cyclists; (2) sometimes this results in a collision.

The spaces left as drivers pass vary; sometimes there is a lot of space, other times there is not. If the whole distribution of gaps shifts by a few centimetres then there is a likelihood that overtaking events which would otherwise have been near-misses become collisions. So I think it does matter, although as ever more research is needed to fully understand the phenomenon.