Invisible Cyclists in Solstice Dark

  • Thread starter Elisa Francesca Roselli
  • Start date



gds wrote:
> Thanks Tony,
> BTW the source of my criticism and frustration with this "debate" is
> that I have had training as an epidemiologist and taught it at the
> graduate level. I am very familiar with epidemiological methods and it
> tis fromthat background that my criticism arises.


So have you read the time series studies because someone with your
background should be able to appreciate that they are probably the best
constructed studies out there and likely to give a reasonable insight
into the answer. They even had control population groups. Yet all your
posts seem to pretend these studies do not exist.

--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham
 
Yes but my conclusion has been stated many times. Those studies do not
yiled the conclusions claimed by folks in this thread. Those time
series studies have not addressed the issue of non reporting and thus
yield no information about what is occurring at a population wide
level.

They are not "best" studies--no matter how many times you say they are.
 
OK Frank, but all your cricisms of my hypothesis dont' matter. Because
from the data presented you can't prove it one way or the other. And
yes it implies that helmets "hurt" half the wearers. That was on
purpose so that you could not accuse me of a pro helmet bias. But
again, the point is that the hypothesis is "possible" and is not at
odds with the study results. And tthus as an alternative hypothesis it
points ut why what you claim to be "findings" are also just a
hypothetical supposition.
I know you folks don't want to let go of this truththat you believe but
it just has not been proven.

So, what would I suggest. I've presented it before but ... oh well here
goes.

I want a prospective population based study. I recognize that the issue
will not command the $$ for world wide study. I'd like to take the
mandatory issue out and concentrate purely on efficacy.

As a first study of its kind I would limit it geographically. Say we
pick the US, but you can choose some other places if you like.
I would then begin stratifiying the US population so that we could do
some proper sampling. Urban, suburban, rural. New England , Mid west,
South east, etc.

We would then draw population based samples (there are plenty of folks
who do this all day including marketing research companies and major
survey research groups at such places as the University of Michigan and
University of Wisconsin) I would then survery the sample population
and determine such things as: do they cycle, if so what types of
cycling activites and how much of each, do they wear a helemt, have
they had an acident, if so did thye get medical help, nature of the
accident, etc. etc. You can add any variables to the list that you
think are important.

And yes this method will not address risk compensation directly but it
would be the first study that actually has the statistics about what is
happening. Depending on what they show folks can decide if further
understanding of "why" is warranted.

I really do not see how you can have the answer to question any other
way.
 
gds wrote:
> OK Frank, but all your cricisms of my hypothesis dont' matter. Because
> from the data presented you can't prove it one way or the other.


.... keeping in mind that "proof" obviously means "proof to the
satisfaction of a particular individual."

> And
> yes it implies that helmets "hurt" half the wearers. That was on
> purpose so that you could not accuse me of a pro helmet bias. But
> again, the point is that the hypothesis is "possible" and is not at
> odds with the study results.


Again, I think any objective person would consider that explanation far
fetched.

But still, look at the results as filtered through your hypothesis!
Neither your hypothetical explanation for the results, nor the much
more likely one, justifies the current enthusiastic, top-priority
promotion of helmets!

> And tthus as an alternative hypothesis it
> points ut why what you claim to be "findings" are also just a
> hypothetical supposition.
> I know you folks don't want to let go of this truththat you believe but
> it just has not been proven.


Frankly, ISTM you've given (obviously without evidence) an alternate
explanation nobody will buy; and if they do buy it, it's practically
equivelent to the much more direct conclusion, that helmets aren't
doing what's claimed.

>
> So, what would I suggest. I've presented it before but ... oh well here
> goes.
>
> I want a prospective population based study. I recognize that the issue
> will not command the $$ for world wide study. I'd like to take the
> mandatory issue out and concentrate purely on efficacy.
>
> As a first study of its kind I would limit it geographically. Say we
> pick the US, but you can choose some other places if you like.
> I would then begin stratifiying the US population so that we could do
> some proper sampling. Urban, suburban, rural. New England , Mid west,
> South east, etc.
>
> We would then draw population based samples (there are plenty of folks
> who do this all day including marketing research companies and major
> survey research groups at such places as the University of Michigan and
> University of Wisconsin) I would then survery the sample population
> and determine such things as: do they cycle, if so what types of
> cycling activites and how much of each, do they wear a helemt, have
> they had an acident, if so did thye get medical help, nature of the
> accident, etc. etc. You can add any variables to the list that you
> think are important.
>
> And yes this method will not address risk compensation directly but it
> would be the first study that actually has the statistics about what is
> happening. Depending on what they show folks can decide if further
> understanding of "why" is warranted.
>
> I really do not see how you can have the answer to question any other
> way.


Well, we may not get an answer that will satisfy _you_ any other way!
Not everyone agrees that your standard for proof is the correct one.

But regarding your proposed study, I think it fails the "reasonable
expense" test. Here's why: Unless you treat every nick, scrape and
bruise as serious enough to catalog, you'll need a huge sample size to
get any statistical validity.

Why? Because during any given year, only a few hundredths of a percent
of the population injure themselves seriously on a bike. And contrary
to popular myth, not every "serious" bike injury is a head injury.

You'll need to contact tens of thousands of people for every person
with a noteworthy brain injury from biking. (And yes, I'm specifically
saying "brain injury" because that's what fearmongers have tied to
bicycling, even though they count cut ears as "head injuries.")

To do this in, say, five areas of the country, in urban, suburban and
rural surroundings, and perhaps to separate kids and adults, and to get
enough "cases" to crunch statistically, you'll need to contact millions
of people.

That's how it seems to me, anyway. IOW, the study that would satisfy
you can't be done - and it can't be done _specifically_ because serious
head injury on a bike is vanishingly rare!

Thus, the very act of examining your proposal should remove the reason
for your proposal. And it removes the justification for the promotion
of bike helmets.

- Frank Krygowski
 
OK fine. But keep up your view and then wonder why all those things are
happening that you don't like.

Don't you see the problem is that even many active long time cyclists
who spent many years not wearing helmets are not convinced by your
argument. For example, you argue that cycling clubs having helmet rules
is not truly driven by insurance. OK we haven't surveyed all the
policies so let's accept that. That must mean that the policies are
being driven by active cyclists commited to encouraging cycling- right?
I mena that would pretty much describe the leadership in cycling clubs.
So, if those folks don't buy your argument but rather have "succombed
to the brain washing of the helmet industry" as you claim; well then
the helmet industry must have a much better argument. Remember these
folks tend to not be new converts to cycling but rather pretty
expereinced cyclists. And clubs run the full range of focii from racing
to recreational riding to touring.

So,you are claiming that all of these experienced, dedicated cycling
advocats have all been fooled and are sadly unable to see the truth of
your argument even with all the data yoou keep presenting.

The real world has voted against your conclusion. That is why most
clubs and certainly almost all organized rides, races, tours, etc
require helmets. They may all be wrong but they almost all seem to be
rejecting your arguments and the research on which it is based. So,
don't you think it may be time for a little critical thinking on your
part to determine why.
 
gds wrote:
> Frank,
>
> I'm having trouble getting more than an abstract on the study you
> reference. And I did find a similar abstract with opposite conclusions
>
> http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V5S-4H9GRM4-1&_user=10&_handle=V-WA-A-W-WE-MsSWYVW-UUW-U-AABEAAZCAZ-AABZZEDBAZ-CDBCEDVEV-WE-U&_fmt=full&_coverDate=01%2F31%2F2006&_rdoc=18&_orig=browse&_srch=%23toc%235794%232006%23999619998%23609292!&_cdi=5794&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=5bfb6bdc1e747467b9f994a0d6b16462
>
>
> And I agree with your comment on practical difficulties organizing a
> double blind study.


I've had time to briefly look over that paper. It's not a very strong
justification for helmet promotion, is it? The correlation between
helmet use and head injury seems weak. And as they say at the end,
"Despite the increase of helmet use, we were not able to show
significant reduction of serious head injury rates over the study
period."

But I was more interested in the total numbers of cases they came up
with. In five years, they had 1,116 cases out of a population of nearly
3 million. That's 0.04% of the population in five years, or less than
0.01% of the population per year. It's not very much different whether
you look at the total population, just kids, or just adults. And
that's not head injuries, that's _all_ serious bike injuries (IOW,
admission to hospital, transfer from acute care center to hospital,
etc.). The percentage with serious head injuries is even less.

That's a small percentage! How much energy should we put into
promoting marginally effective plastic hats - and into scaring people
away from bicycling by that promotion?

- Frank Krygowski
 
gds wrote:

> The real world has voted against your conclusion. That is why most
> clubs and certainly almost all organized rides, races, tours, etc
> require helmets. They may all be wrong but they almost all seem to be
> rejecting your arguments and the research on which it is based.


Actually, the helmet requirements often come from the insurance
companies for the clubs and events. It's the insurance companies that
are trying to reduce their exposure. Clearly the insurance companies
have voted against whoever's conclusion you were referring to.

> So,
> don't you think it may be time for a little critical thinking on your
> part to determine why.


Unlikely to happen.

As a side note, please include at least a snip of who and what you're
replying to when you do follow-ups.
 
At the 95% CI the study reported P values that were worth looking at-
including the pre and post legislation rates of serious injury and the
injury raes for helmet users vs non users.. The rates for those that
used helmets were roughly 1/2 the rate for non helmet users.

I think that halving the risk would be considered successful by most
folks.

BTW I think you need to read some more on sample size and statistical
inference. If properly constructed pretty small samples can yield lots
of statistical reliability and validity. All of your comments about
needing sample sizes inthe millions just isn't so.
 
gds wrote:
> Yes but my conclusion has been stated many times. Those studies do not
> yiled the conclusions claimed by folks in this thread. Those time
> series studies have not addressed the issue of non reporting and thus
> yield no information about what is occurring at a population wide
> level.
>
> They are not "best" studies--no matter how many times you say they are.
>


They are the best - "best" is a comparative not an absolute.

But the point of a time series study like that is that there is a step
change in one factor, helmet wearing, during which there is minimal
change in other factors. So non-reporting will have occurred equally
both before and after the step change. What they show very clearly is
that a step change doubling in helmet wearing produced no measurable
change in the head injury rate. Now if TR&T were correct doubling the
helmet wearing to approx 100% should have reduced head injuries
requiring hospital admission to about a quarter of their previous level.
But it didn't. Unless you have good evidence that helmet laws forced
people to report injuries of a triviality they had never reported before
at a level that exactly cancelled out the reduction in serious head
injuries, no more, no less, the inescapable conclusion is that helmets
do not appear to provide the protection claimed for them.

--
Tony

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

> BTW I think you need to read some more on sample size and statistical
> inference. If properly constructed pretty small samples can yield lots
> of statistical reliability and validity. All of your comments about
> needing sample sizes inthe millions just isn't so.


Sigh, this seems to be typical Usenet problem. I remember explaining how
a sample size of 0.04% was actually a very large sample (this was for
a Consumer Reports survey, with a sample size of 31,000 cellular phone
subscribers out of about 70 million subscribers total) yielding a margin
of error of less than 1%. It's very hard for some people to believe that
statistical sampling is actually valid, unless they have been trained in it.

There is a margin of error calculator at
"http://americanresearchgroup.com/moe.html" which is fun to play with.
 
[email protected] wrote:

> In the past, I've posted "per hour" data from at least five different
> countries. I suppose I can do it again.


Oh, please do. Then we can all join in making fun of your
'data' which you post without ANY sort of methodology
or background material, then demand that we accept.
Please Frank, stop this ridiculous charade. It's
embarrassing.

> On the one hand, it's true that there are no hour meters on bicycles.
> Nonetheless, professionals dealing with public safety in many different
> countries around the globe have used various techniques to get their
> best estimates of hours of use.


'Various techniques.' That's rich. Does one of these
various techniques involve drunken magazine interns
hurling darts at a piece of paper with random numbers
on it?

> They've published estimates of
> fatalities per hour for cycling, and the values tend to be in broad
> agreement.


'Broad agreement.' No, they're not. Please post those
numbers again so you can be properly hoisted by
your own pittard.

> In summary, we have professionals around the globe doing what they need
> to do;


Any true professionals who are involved in this sort
of activity will tell you that results from it are not
scientific, but are rough estimates at best, and should
not be abused as 'data' in any serious discussion.

> and we have a bike messenger saying it's impossible to do it.


Nonsense. The bicycle messenger himself has offered
per-hour ESTIMATES in this very thread.

> Robert, if a woman with a PhD in public health, whose job was
> collecting transportation data, told you it's impossible to do a track
> stand, would you believe her?
>
> Or would you say "I'm the expert cyclist. Stick to your own field." ?


No. I would SHOW her exactly how it could be done.

Got it?

It's not a hard concept.

Robert
 
SMS wrote:

> There is a margin of error calculator at
> "http://americanresearchgroup.com/moe.html" which is fun to play with.


Yes. Taking a population of 31 million and a sample size of 10 thousand
gives a margin of error of 0.98. Wow, pretty good!

Except that you are looking at a head injury rate of 0.01% so any
change is going to be insignificant.

Ten times the head injury rate is going to be still one tenth of the
sampling error.

If you are going to try to play stats, at least do it right.

...d
 
[email protected] wrote:

> 'Various techniques.' That's rich. Does one of these
> various techniques involve drunken magazine interns
> hurling darts at a piece of paper with random numbers
> on it?


Surveys show that this to be the case. There is broad agreement about it.
 
gds wrote:
> At the 95% CI the study reported P values that were worth looking at-
> including the pre and post legislation rates of serious injury and the
> injury raes for helmet users vs non users.. The rates for those that
> used helmets were roughly 1/2 the rate for non helmet users.
>
> I think that halving the risk would be considered successful by most
> folks.
>


So why do you think the graphs for head injuries in adults and children
(Figs 1 & 2) only move within normal statistical fluctuations, with
significant changes in helmet wearing levels. Indeed in the first year
of Fig 1, helmet wearing quadruples but the head injuries go up, not
down, although I doubt the increase in injuries is more than statistical
fluctuations in the small sample. Meanwhile in Fig 2 helmet wearing
goes from 0 to 40% in 3 years yet head injuries stay within the sample
natural noise variations of 25 +/- SQRT25.

Why do you think that is? Could it be because their conclusion is
"Despite the increase of helmet use, we were not able to show
significant reduction of serious head injury rates over the study period."

But how do you explain the complete lack of any effect of major
increases in helmet wearing on head injury rates?

--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham
 
The same way I explain the problem with all the other studies. The
study looked at reported injury rates rather than what was going on in
the poulation--again ignoring non accidents and non reported accident
which might (or might not) be impacted by helmet use.

So, how do you calculate a rate without knowing the numerator (Total
cycling injuries reported and non reported) or the denominator (the
corresponding population of cyclists)?

My point in the reference was in the halving of the reported accidents
for helmet users. While not statistically signifigant it does point to
what I think is a reason to look further at what is happening overall.

I'll go back to my earlier response. If the reality is that there is no
problem, i.e. the risk of injury is so miniscule that it is not worth
any worry whatso ever. And even if one did worry helmets had no impact
on injury severity anyway. Then why are so many experienced cycling
advocats, folks with no financial interest to either insurance or
helmet industries instituting and supporting helmet use? It seems that
large numbers of experienced cyclists are rejecting the research you
proffer. Why?

OK, then Frank say go look at the non cycling club cyclists- they are
they ones rejecting helmets. Well, it I don't know what to make of
that. Is the argument that all of these dedicated high mileage
recreational and racing cyclist should be discounted and we should only
listen to commuting cyclists and, of course, all those non categorized
riders that ride the wrong way at night without lights. But yet it is
ok to reject the behaviour of all those fast guys on nice bikes because
they wear lycra.
 
gds wrote:

> OK, then Frank say go look at the non cycling club cyclists- they are
> they ones rejecting helmets.


Why do you think this? It varies by location obviously, but don't
generalize. Helmets are the rule, rather than the exception, among
commuter cyclists in many areas.

Please don't extrapolate one commuter rejecting helmets, into anything
like a majority of non-cycling club cyclists rejecting helmets.
 
SMS wrote:
> gds wrote:
>
> > OK, then Frank say go look at the non cycling club cyclists- they are
> > they ones rejecting helmets.

>
> Why do you think this? It varies by location obviously, but don't
> generalize. Helmets are the rule, rather than the exception, among
> commuter cyclists in many areas.
>
> Please don't extrapolate one commuter rejecting helmets, into anything
> like a majority of non-cycling club cyclists rejecting helmets.


I didn't extrapolate it, I was simply quoting Frank.
 
On 9 Jan 2006 18:20:12 -0800, [email protected] said in
<[email protected]>:

>> Apart from the ones that don't, such as those Frank has posted.

>Which ones are those?


That explains it: your perceptual filters prevent you from seeing
anything which might conflict with your macho bike messenger view of
cycling as a dangerous activity. This explains why arguing with you
has always been a waste of time.

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

"To every complex problem there is a solution which is
simple, neat and wrong" - HL Mencken
 
On Tue, 10 Jan 2006 12:41:49 +0000 (UTC), Chris Malcolm
<[email protected]> said in <[email protected]>:

>in my case it's clear that gardening is well over 100 times more
>injurious per hour than cycling.


Massively more so in my case: I have needed surgery under general
anaesthetic to fix one gardening injury, no cycling crash has ever
been that bad!

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

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

"gds" <[email protected]> wrote in message
news:[email protected]...
>
> SMS wrote:
>> gds wrote:
>>
>> > OK, then Frank say go look at the non cycling club cyclists- they are
>> > they ones rejecting helmets.

>>
>> Why do you think this? It varies by location obviously, but don't
>> generalize. Helmets are the rule, rather than the exception, among
>> commuter cyclists in many areas.
>>
>> Please don't extrapolate one commuter rejecting helmets, into anything
>> like a majority of non-cycling club cyclists rejecting helmets.

>
> I didn't extrapolate it, I was simply quoting Frank.
>


Hey you guys why not let it drop! You are very boring and cross posting to
boot

--
Trevor A Panther
In South Yorkshire,
England, United Kingdom