Thursday, September 14, 2006

I'll Take My Chances With The Helmet

by Tom Bozzo

Via Mark Thoma, Andrew Gelman discusses a study (summarized here) by a U.K. researcher suggesting that bike helmets may be dangerous to cyclists because motorists drive more agressively around cyclists they perceive to be experienced.

Ian Walker of the University of Bath recorded the distance at which motor vehicles passed him for a large number of interactions during rides with and without a helmet; he found that the average distance was 3.3 inches less for the helmeted encounters; moreover, he was hit twice while riding the helmeted portion of the experiment. That Walker's accidents didn't seem to materially impede his work may vouch for the efficacy of helmets conditional upon an accident; he understandably didn't partake in helmetless 'control' accidents.

Much of the discussion, explicit (at Gelman's) or implicit (at Marginal Revolution) is on the driver psychology effect. But while the 3.3 inches is characterized as (at least statistically) significant, the large fraction of the variance in distance not explained either by helmet-wearing or the distance Dr. Walker rode from the edge of the roadway suggests clearer implications for traffic engineering.

From my bike commuter's perspective, much of what makes Madison a cyclist's paradise — at least when the temperature is above freezing — is the combination of dedicated trails and other cycling routes with marked bike lanes or other accommodations for cyclists (e.g., wide shoulders). Thankful as I am to those motorists who are kind enough to share the road, I see my welfare as being improved when there's less need to share in the first place.

In the former case, there just aren't cars around to worry about, so the principal hazards are other cyclists and careless (i.e., absorbed in mobile telephony) peds; potential collision speeds and energies would seem to be well in the range where preventing inadvertent skull-cracking is both within the helmet's repertoire and addresses the major health risk — unlucky landings aside.

In the latter, 3.3 inches' difference in approach distance would be harder to notice, as I've observed typical passing distances of several feet, though the movement in the tail of the distribution may lead to a large relative increase in a small risk of collision, a la Walker's findings. (I may not be Lycra-clad, but my helmet and pannier would presumably peg me as a reasonably serious commuter under the driver psychology hypothesis.) Otherwise, I doubt that the helmet would have much marginal effect on motorists barreling down bike lanes.

The accommodations are particularly important for interactions with heavy vehicles. Walker's two accidents were with a bus and a truck. Both buses and heavy trucks are reported as passing several inches closer, on average, than light vehicles. Offhand, vehicle width differentials would seem to account for much of the difference. On narrower streets, those few inches could easily make the difference between a close shave and a trip into a ditch or worse.
There is clearly a cultural effect, for the lack of a better term. In some parts of the US (i.e., the Bay Area, a cycling haven), it's the lack of a helmet, more than the presence of one, that suggests inexperience.

In practical terms, then, the risk-averse rider would first need to figure out whether donning a helmet would increase or decrease the chances of being taken for an experienced rider in a given locale. Cross-country bike tours would become rather tedious.
You have a loose italic somewhere in that post, by the way.

I'll stick with the helmet, and concur with Kim that—assuming the rider is over eight, or maybe ten—helmets are usually a sign of experience.

When I was riding regularly (never more than 120-150 mpw), a lot of that was done on 9-W or in Central Park, neither of which has enough space on the side of the road to give an extra 3.3 inches to someone, even if they were wearing a wig.

But there is a reason that NYC always rates at top of the Least Friendly Cities list.
If it wasn't clear from the post, the driver behavior theory is that the helmet *is* viewed as a sign of experience, and such riders are given less leeway.

My sense is that drivers who actually notice cyclists give more leeway; my closest approaches have been by cell phone talkers exercising predictably poor control over their vehicles.

(BTW, Ken, I hadn't noticed any formatting issues on either the home or work computers.)
ooops. Clearly I didn't read carefully enough. (I blame my day's work, which has actually been a week's work, which is to explain discrepant results in a variance decomposition of hourly earnings. Simple stuff, really, but trying to second-guess how someone got a result, especially when you don't have a copy of their program, is mind-numbingly frustrating stuff.)
Kim, I understand the frustration. I've been dealing with some Stata code that's relatively artful programming but also so modular that you'd think it was designed to frustrate tracing dependencies. (And since it's documentation filed by an opposing witness, I actually do think that.)
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?