Re: Some traffic stats

On 3/14/2024 1:53 AM, Jeff Liebermann wrote:

On Thu, 14 Mar 2024 00:00:45 -0400, Frank Krygowski
<frkrygow@sbcglobal.net> wrote:
 

On 3/13/2024 4:06 PM, Zen Cycle wrote:

On 3/13/2024 3:44 PM, Frank Krygowski wrote:

On 3/13/2024 11:34 AM, AMuzi wrote:

https://www.cityofmadison.com/police/newsroom/incidentreports/incident.cfm?id=30855

>
  From that site:
>
* On average, every day, twenty pedestrians are killed by a moving
vehicle in the United States.
>
* Approximately 76,000 pedestrians and 47,000 bicyclists are injured
in roadway crashes annually in the United States.
>
I'll note that the figures for pedestrians are far worse than for
bicyclists. Yet the general public thinks of bicycling as much more
dangerous than walking.

>
I think if you compare injuries per participants or injuries per miles
traveled, you'll see they're probably correct.

>
That's far from certain.
>
Powell et. al., “Injury Rates from Walking, Gardening, Weightlifting,
Outdoor Bicycling and Aerobics”, Medicine & Science in Sports &
Exercise, 1998, Vol. 30 pp. 1246-9 polled over 5000 people who had
chosen at least one of those activities for exercise. One question was
whether the participant had incurred an injury during the previous month.
>
The results:
Weightlifting: 2.4% of participants injured
Gardening or yard work: 1.6%
Aerobic Dance: 1.4%
Walking for exercise: 1.4%
Outdoor bicycling: 0.9%

 Such surveys require a control group to be valid.  I suspect that if
someone was able to do NOTHING for one month, they would still have
been involved in some kind of accident.

Meh...I don't think it's all that relevant for pure statistical analysis. Do you really need a control group to compare injury rates per mile for walking vs cycling? I don't think so.

 

And while injuries =/= fatalities, Dr. John Pucher of Rutgers has
published (in "Making Walking and Cycling Safer: Lessons from Europe")
an estimate from U.S. data that bicyclists suffer 109 fatalities per
billion km ridden.  Pedestrians suffer 362 fatalities per billion km,
three times as bad!
>
Pucher's number works out to 5.7 million miles ridden per fatality for
cyclists, 1.7 million miles walked per fatality for pedestrians. And
Pucher's later work, as well as other sources, show he greatly
overstated the bicycling risk. It's now widely accepted that Americans
ride over ten million miles between fatalities.

 Americans ride over 10 million miles between fatalities?  Most
Americans don't ride after their first fatality.  The value of
exercise after death has been greatly overrated.
 

British data for decades has consistently found more pedestrian
fatalities per mile traveled than bicycling fatalities per mile. AFAIK,
there have been only a couple years in the past 20 where the reverse was
true. I've also seen Australian data showing the same result.
>
In any case, for most Americans the far bigger danger is sitting on the
couch.

 If someone asked you "what is your favorite sport and how many times
have you been injured in the previous month", would you produce an
accurate number, or would you minimize the number of injuries?  My
past experience working with such surveys suggests that most people
would not admit to an injury.

True, self-reporting surveys are always suspect. One would suspect just the participation rates are highly suspect, as noted in the massive differences between the referenced NHIS study and the ICARIS study.

I also find it odd that the survey
would ask if "the participant had incurred an injury" instead of
asking how many injuries.  Why only one month?  Were they worried that
if they extended the time period to one year, a much larger percentage
would probably have been injured at least once.  Also, was the month
the same for everyone in the survey?  I suspect not or the question
would have been phrased differently. 

That was all detailed in the body of the study. They discussed specifically why only 30 days, and FYI, it was the "past 30 days".

For cycling, if they selected a
month which has a high accident rate, typically when everyone is
cycling, the survey results would have been very different had they
selected a month with a low accident rate.  Time of day when riding
would also have had a huge effect on the accident rate.  Were riders
who died included in the survey?  Traditionally, the dead do not
answer survey questions.

The dead aren't liekly to be interviewed at all. The ICARIS study referred to in Powell et al was based on USDOT data and included reported deaths.

 How far did they ride in the month of the
survey?  If it was rather long distances, the chances of having an
accident would be much higher than if they rode short distances.

True, the Powell study didn't address that as part of the survey, which is one reason I noted to frank that it leaves me unconvinced.

       

--
Add xx to reply