Invisible Cyclists in Solstice Dark

In uk.rec.cycling SMS twisted the electrons to say:
Hmm, I see that you claimed this in a letter, but I never saw them make
that claim.

He claimed it in a letter using the figures that they provided in their
report. If you beleive him to be in error, then should you should show
how and why. You're never going to get full marks in the exam if you
don't show your working out! grins
--
These opinions might not even be mine ...
Let alone connected with my employer ...

On Fri, 06 Jan 2006 14:35:59 -0800, SMS
said in :

This one, where Cook and Sheihk (authors of the book "Basic skills in
statistics") claimed that helmets prevented 186% of accidents:

Hmm, I see that you claimed this in a letter, but I never saw them make
that claim.

It's very simple, just do the maths correcting their schoolboy error.

Guy
--
http://www.chapmancentral.co.uk

"To every complex problem there is a solution which is
simple, neat and wrong" - HL Mencken

On Thu, 05 Jan 2006 23:37:08 GMT, Mike Rice said in
:

Yes, Steve reckons that the tiny side-facing LEDs in the TL-1000 will
provide a level of protection against side impacts massively greater
than that which the (far brighter) rear-facing LEDs in cheaper rear
lights afford against rear impact. I think in ScharfWorld rear
impacts are more common than they are on planet earth :-)

Guy I generally agree with you, but I use a TL 1000 and really like
it. One problem I had with my older LED rear light was that
occassionally the battery would jostle loose and the light would stop
functining. Because it was rear firing I would not be aware that it
was out until arriving at my destination and dismounting.

I have an LD1000 on the Brom. But the LEDs are no more visible from
the sides than the main LEDs on my several B&M 4D TopLites.

More importantly, we both know that the side LEDs are not large.
Scharf, in a previous post, was denouncing competing LED lights as
being insufficiently bright, but suggesting that the side LEDs in the
TL-1000 could prevent side impacts - a logically inconsistent
position, as I'm sure you'll agree.

My biggest problem with Scharf on the issue of lights is this: he
asserts that *only* very bright lights are adequate. Since I know of
no evidence that lights make any difference at all, however much I
spend on them (which is a lot), I don't think Scharf's assertion holds
water. I am not alone in this :-)

Of course I think the best method of preventing side impact crashes is
to ride defensively and never trust a driver with your safety.

Quite.

Guy
--
http://www.chapmancentral.co.uk

"To every complex problem there is a solution which is
simple, neat and wrong" - HL Mencken

SMS wrote:

James Annan wrote:

SMS wrote:

Alan Braggins wrote:

Ah, but we know that in some studies, helmets protect against more than
100% of injuries.



Huh? Which studies were those?


This one, where Cook and Sheihk (authors of the book "Basic skills in
statistics") claimed that helmets prevented 186% of accidents:


Hmm, I see that you claimed this in a letter, but I never saw them make
that claim.

It's the bit where they write:

"Assuming that the difference between cyclists and pedestrians (3.6%) is
attributable to helmet wearing, and with helmet wearing having increased
by 5.8%, we estimate that helmets prevent [erroneous calculation] of
serious head injuries."

http://ip.bmjjournals.com/cgi/content/full/9/3/266

Using their assumptions, the correct calculation gives 186%. They didn't
quote that number because they got their sums wrong - a point which they
did explicitly acknowledge in their reply to my letter.

Their data actually prove clearly that there are large uncontrolled
factors other than helmet-wearing which dominate the observed changes in
head injury rate. The authors make no attempt to investigate or quantify
any of these effects. They assumed a priori that helmets would be the
biggest effect, and because they got their sums wrong, they never
realised that the data contradict this belief.

James
--
James Annan
see web pages for email
http://www.ne.jp/asahi/julesandjames/home/
http://julesandjames.blogspot.com/

Just zis Guy, you know? wrote:
Which is, presumably, why you keep banging on about it, since the
measures for which we have a reasonable estimate of exposure all show
that cycling is actually quite safe,

'Actually quite safe.' You mean 30 times more likely
to cause injury per mile than driving actually quite
safe (Moritz)?

and we all know how that idea
offends your macho bike messenger pride :-)

Indeed, my macho bike messenger pride is offended.

Robert

wrote:
The following is not a scientific paper, but it does refer to the rise
in head injury rates in the US, despite the growing popularity of
helmets.

http://www.magma.ca/~ocbc/nytimes.html

- Frank Krygowski

The main points in this New York Times 2001 article
http://www.magma.ca/~ocbc/nytimes.html (cited by Frank Krygowski)

- I've summarised here..
http://www.beseenonabike.com/bsoab/b...facts.htm#head

- the rate of head injuries per active cyclist has increased 51 percent
just as bicycle helmets have become widespread. Why?

Reasons cited in the above article, for this increase in head injuries
with increased helmet use a-

- Helmets are worn incorrectly, for example, helmet worn to far back
leaving forehead unprotected

- Daredevel Effect: with a helmet, cyclists may feel an inflated sense
of security, may ride faster and take more risks. A parallel situation
is anti-lock brakes, when introduced ABS brakes were supposed to reduce
accidents, but drivers realized their brakes were more effective & they
started driving faster - (some) accident rates rose.

- to reduce the Daredevil Effect "you would be well advised to wear a
helmet provided you could persuade yourself it is of little use"

And

- "patients who were wearing helmets when they were injured are much
better off than those who were not"
- "Bicycle helmets are the best technology we have for protecting the
brain, the helmets serve the function of an air bag."

Chris Street
www.BeSeenOnABike.com

gds wrote:
Clive George wrote:
Um - no?

If the total number of injuries severe enough to require hospital treatment
remains the same regardless of the fact that more helmets are being worn,
that indicates to me that wearing a helmet doesn't affect my chance of
getting seriously hurt in x years/km of cycling - the measure I mentioned
above as being the important one.

The problem with the hypothesis you present above - that helmets make some
accidents minor and hence unreported - is that there doesn't appear to be a
corresponding reduction in the serious injuries suffered. If the number of
accidents remained the same, the helmets aren't working. If the helmets are
working, the number of accidents has gone up. Either way the chance of
getting a serious injury - which is as I may have said before, the only
important factor here - is not reduced.


Not quite! Despite prsentations on the effect of mandatory helmet laws
and decreased ridership in some areas there is also separate data
showing that, for example, in the US bicycle ridership has increased
during the same time that helmet sales and use increased (no mandatory
effect). Call it the Lance effect or whatever ( a few years ago it was
the Greg Lemond effect). So, we also need to understand the
denominator. If bicycle miles and/or time increased during the period
in which helmet use increased you could have the result that the same
number of serious accidents represented a smaller accident rate and you
could hypothesize that the rate reduction was due to helmet use.

Again, I'm not arguing that that is so; merely that without
understanding all of these variables you can't possible understand the
totality of what is happening.

Yes you are arguing that it is so when it isn't. The denominator is
understood. It is measured by a number of different methods, none of
them perfect. But if you make a measurement ten different ways and they
all come out with roughly the same number you have a reasonable level of
confidence that it right and certainly it is more likely to be right
than wrong. In which case the conclusions on pre mile risk that helmets
do not help is more likely to be right than wrong - something you seem
to be desperate to deny


It seems to me that lots of you folks seem to oppose helmets stemming
out the mandatory situation in some jurisdictions. I can sympathize
with that sentiment. But I don't think the way to address it is to
pretend you have good statistical evidence when you don't. I'm all for
letting folks decide for themselves-- and Ijst wish everyone had much
better data on which to base that decision.


The data is of varying quality - there are some very competent studies
and some crap ones like TR&T. However the credible statistical evidence
is overwhelmingly in support of helmets at best not helping and possibly
making the situation worse. By credible I rule out studies that show
that helmets prevent 75% of leg injuries or 186% of head injuries. The
most statistically robust studies are the time series studies where a
step change in helmet use took place due to MHLs and confounding factors
can be largely ruled out. When all the different studies in different
countries by different methods tell you one thing the probability is
that that thing is more likely to be right than wrong. Something again
you seem desperate to deny.

--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham

wrote:
SMS wrote:
One of the papers, that someone posted a reference to, switched between
per hour and per mile as it suited the authors needs. It was filled with
unreferenced weasel statements, such as "surveys show," but of course
with no reference to any such survey.

Steven, can you just _once_ give us a proper citation to an article
you're talking about?? Just once??


I've noticed a pattern in this thread which is whenever Scharf the Larf
gets to a situation is sub-thread where his position is untenable,
instead of accepting it he suddenly evaporates from that sub-thread and
pops up elsewhere with a different argument. I notice he has still not
dared take the test I set him to tell me where on two graphs of head
injury numbers did an MHL double helmet use.

--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham

wrote:

And only then will Robert and Steven be convinced that cycling (and
walking) are acceptably safe.

Right, guys? ;-)


Right, which is probably a good point to draw a line under this thread
given that the one thing that has been determined is that Robert and
Steven are not interested in any answers unless it supports their faith
in helmets.


--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham

["Followup-To:" header set to uk.rec.cycling.]
On 6 Jan 2006 15:32:51 -0800,
wrote:
Just zis Guy, you know? wrote:
Which is, presumably, why you keep banging on about it, since the
measures for which we have a reasonable estimate of exposure all show
that cycling is actually quite safe,

'Actually quite safe.' You mean 30 times more likely
to cause injury per mile than driving actually quite
safe (Moritz)?


You've got this backwards (at least in the UK)

Fatal injuries involving cars but not cycles (~3500)
Fatal injuries involving cycles but not cars (~10 but most of these are
offroad, on the road/pavement the number is ~1)

pkm for cars ~500bn
pkm for cycles ~5bn

cars - 7 fatalities per bn pkm
cycles - 2 (0.2) fatalities per bn pkm


If you ignore injuries to pedestrians it's still in favour of cyclists
even though I only know of one case where a pedestrian (crossing a
motorway) killed a driver.

fatal injuries to car occupants in single car or car/car accidents
(~2500)

fatal injuries to cyclists in single cycle or cycle/cycle accidents (~10
but most are off road. On road the number is 1)

cars - 5 fatalities per bn pkm
cycles - 2 (0.2) fatalities per bn pkm


Tim.


--
God said, "div D = rho, div B = 0, curl E = - @B/@t, curl H = J + @D/@t,"
and there was light.

http://tjw.hn.org/ http://www.locofungus.btinternet.co.uk/