[SystemSafety] Bicycle Helmets

[Peter Bernard Ladkin]

On 2014-10-13 13:58 , Peter Bishop wrote:
> If only 1 in 7 cyclists wear a helmet, and we assume helmets have no effect, we would expect 117/7 ~
> 17 helmet-wearing cyclists in 117 fatal accidents.
> 
> At first sight the observation there are only 6 fatal accidents with helmets indicates a halving of
> the risk of death to the cyclist, 

I am not sure about the number of 1 in 7 being anything meaningful. I doubt it's a good proxy for
exposure, which is what needs to be counted. There is a difference between counting cyclists who
wear a helmet (sometimes) and counting exposure: the total time spent cycling with helmets versus
the total time spent cycling without.

And then there is location. In Bielefeld city, there are now a lot of bikes around and I'd say at
least half the commuters wear helmets. Those are the most exposed (in the technical sense). So I'd
say the exposure of helmet-wearing people could well be 1/2 or more. Given that there are two or
three fatal bicycle accidents in Bielefeld per year, and Bielefeld is the 20th or so largest city in
Germany, it could well be that those accidents represent 5 years of the 40 largest cities, in which
case the helmet-wearing exposure would be different from one in seven if Bielefeld is any guide.

Casual cyclists might not wear a helmet, are less exposed and usually significantly slower but
equally less experienced. Weekend rural cyclists are a different matter from Bielefeld-city
cyclists, except of course for city-dwellers on tour. But all hobby racers now wear helmets, and
around my village they account for a half or more of weekend traffic.

Doing it coarsely doesn't necessary yield any info. Say I divide the total population of Germany by
the number of muggings in a year to get the number N and then tell my neighbor chosen at random that
there is a one-in-N chance she's going to hit somebody badly enough this year for himher to seek
medical attention, while stealing hisher wallet (I am speaking here outside of the marriage :-) ).
For some of my neighbors, N is infinity. For others, well, I've been here eight years and it hasn't
happened yet so I assume it's a little less than one-in-eight :-)

Exposure confounds a lot of our favorite comparisons. Almost all the exposure to the main hazards in
commercial flight resides in about 20 minutes per flight, independent of journey time, whereas the
exposure to hazards in driving is roughly uniform throughout the journey.

>.. with the small numbers involved I am not sure it is
> statistically significant

Indeed so!

PBL

Prof. Peter Bernard Ladkin, Faculty of Technology, University of Bielefeld, 33594 Bielefeld, Germany
Tel+msg +49 (0)521 880 7319  www.rvs.uni-bielefeld.de