Saturday, April 19, 2008

Conspicuity and Crashes

A study of bicycle crashes here:

and one of motorcycle crashes here:

both show that riders who wear high visibility clothing have much lower crash rates than their more subtly-dressed peers. One problem with such studies is that human beings are complicated creatures. For example, the cycling study states:
Risk-taking behavior is a potential confounder in all observational studies of injury and is difficult to measure reliably because of the likelihood that participants will provide socially desirable responses. "Reverse causation" can also complicate cross-sectional analyses, if outcomes influence exposures. In this study, such a mechanism would be expected to reduce associations rather than inflate them—for example, a bicycle crash may prompt a participant to use fluorescent clothing.
In other words, if you're prone to be a risky rider you probably won't identify yourself as such and if you've had an accident you might load up on safety gear. And I personally think one huge confounder might be this: folks who think a lot about safe riding will probably be far more prone to be swayed by safety considerations in the purchase the fluorescent gear. Riders may not be safer because they are wearing the bright colors, they may be wearing bright colors because they are more safety conscious riders.

While it is very hard to allow for all the confounding factors, both studies suggest that brightly dressed riders have lower crash rates. In any case, I figure my habbit of wearing a bright yellow vest while riding can't hurt.
Post a Comment