(I'm sorry to be hijacking a UK newsgroup to discuss a German newspaper
article. If anyone reads German and is interested, it can be found here:
http://www.sueddeutsche.de/automobil/artikel/529/81448/
)
mark ([email hidden]) wrote:
: Marianne Promberger wrote:
: > What is the base rate over which a 20 percent increase is observed?
: > Is it 1,000 or is it 10?
: The article mentioned 575 killed in 2005, so presumably 460 +/- in 2004.
: Not sure if that's a big enough base to rule out the occasional
: statistical blip.
But the base is what you need if you want to use that article to say
anything about UK vs Germany comparison. I fully agree with your
conclusion below (the grass not being universally greener on either
side); I just don't think there's anything in that article to support
that conclusion. Except maybe that even the German press likes idiotic
populist scaremongering stories about cycling from time to time.
: > No data whatsoever in that SZ article about exposure/
: > cyclist mileage trends.
: >
: > Year to year fluctuations are quite likely random.
: >
: > That SZ article drove me up the wall. Wonderful scaremongering. Title
: > was "Dangerous Cycling -- Death on Two Wheels". And a spokesperson for
: > ADFC (roughly, German equivalent of CTC) was plugged as saying "A cycle
: > helmet can reduce a head injury by 80 percent." Both the ADFC and the
: > VCD, another organisation lobbying for changes in transport structure
: > (sorry to be so vague, I really don't know them too well) made the data
: > out to be compelling evidence for a serious problem. No-one asked the
: > question whether cycle use had increased, and no-one thought about the
: > fact that we really can't tell anything from comparing relatively small
: > numbers for two years.
: >
: The 80% figure is indeed [censored]. In their defense, neither group seemed to
: think that this was grounds for mandatory helmet laws.
Well, ... it is very naive to think that you can make blanket
statements of miraculous safety devices and pretend it won't have
consequences, if not in laws, then at least in social pressure and in
insurance companies refusing to pay in case of an accident. I would
have expected a dedicated *spokesperson* of a group that pretends to be
the cyclists' lobby to be a little more careful in her wording.
: The article made
: the very interesting point that most of the rise in fatalities occurred
: among children and the elderly, and that these were two groups that
: could benefit more from helmet use than the general cycling population.
Yes, and regarding the bigger increase that very interesting point was
based on looking post-hoc at some subgroups of the data, which is
tricky. For the elderly, it says that the numbers went up by more than
25 percent. No mention whether that is in any way significantly
different from the 20 percent overall. For kids, it goes up by 78
percent, but to a total of 41, so a very small subgroup sample size.
It is probably true that kids and the elderly could benefit most from
wearing helmets when cycling, but they would benefit most from all
kinds of safety equipment in all of their activities, cf. thudguard.
: As for cycle use increasing, my understanding is that cycle use has been
: increasing over the years in Germany, but there hasn't really been a
: sudden surge in cycling activity (like London after the tube bombings)
: that would produce a 20% increase in fatalities.
But cycle use may just fluctuate form year to year. Good summer vs bad
summer etc.
When you think about cyclist safety, you care about the absolute
numbers related to exposure, and when deciding about policy, you care
about the *long-term* trend, not year to year fluctuations. Just as
when looking for long-term stock investment you don't jump to
conclusions based on one year's performance of some company.
There was an article in Radzeit 4/2005, the magazine of ADFC Berlin,
that between 1978 and 2004, total numbers of cyclist deaths decreased by
70 percent, (as opposed to 66 percent for all traffic), while cyclist
percentage of all traffic had gone up from 4 percent to 12 percent
during that time. Of course, we have the same problem here, in that 1978
may have been an exceptionally bad year, and 2004 and exceptionally good
year. Indeed, Radzeit notes 1978 as the "all time high" in death rates
for cyclists. 2004 being, by chance, a benign year would make it more
likely for cyclist deaths to bounce up in 2005. Regression to the mean.
While this spurious picking of two years makes this trend shaky as well,
the overall trend seems to be quite positive, and especially, it is a
trend for an activity with a good absolute safety record to begin with,
a message that never gets mentioned in the SZ article.
: I've enjoyed my (very limited) experience cycling around London, and I
: had a great time cycling around Scotland and Wales (and even parts of
: England). I posted the reference to the SZ article because I get tired
: sometimes of people proclaiming loudly that the grass is universally
: greener on the other side of the fence, which it seldom is.
I completely agree with that assessment.
: And yes,
: there is a lot of good beer to be drunk in Germany.
With that one as well, and I have a pretty large sample size over a
long period of time to back it up 😉
Cheers
m.
