[extropy-chat] Air / Car Accident rates

[BillK]

On Wed, 19 Jan 2005 15:24:54 -0500, Eliezer Yudkowsky wrote:
> Experiments in cognitive psychology show that people fear air travel more
> than car travel.  Why?  Because airplane accidents are more frequently
> reported in the media.  People's fear of a particular hazard is, generally,
> directly proportional to how often that hazard is reported in the media.
> Rare hazards that kill lots of people, like airplane crashes, tend to be
> reported in the media often; common small hazards go relatively unnoticed.
> This bias is only one of many experiments where people have been observed
> reason, not from the actual statistical frequencies of cases, but from the
> most attention-grabbing cases or most commonly reported cases.

In the midst of an excellent post on Bayes and reality, Eliezer
slipped in the above reference to air travel safety.  Now I am sure
that the airline industry publicity machine claims that air travel is
much safer than car travel, but you have to look at their claims
carefully.

Obviously, in actual numbers many more people are killed on the roads
than in air crashes. But many more people are on the roads than are in
aircraft.

The air industry would like you to compare accidents per mile
travelled, which is logically silly, because only about 10% of air
accidents occur in the cruise phase of the flight that covers great
distances. If you call them on this, their fallback position is to
compare accidents per 100,000 hours of travel time. This makes their
accident rate about 20 to 40 times worse (!), but is still grossly
flattering, due to the long distance cruise phase where nothing much
happens (and even the pilots are asleep).

The best measure is to compare fatalities per journey, but you will
find it difficult to get these figures.

<http://www.planecrashinfo.com/cause.htm>
provides some interesting tables and graphs.

Using the flattering fatalities per 100,000 hours rate they conclude: 
It is 4 times safer to fly on a airliner than to drive.
It is 4 times safer to drive than to fly on a scheduled commuter airline.
It is 12 times safer to drive than fly on a non-scheduled commuter
plane (air taxi on demand).
It is 26 times safer to drive than fly in a general aviation private plane.

Method note: To compare airplane fatality rates to automobile fatality
rates one must use the same measurements. Unfortunately airplane
fatality rates are expressed in number of fatalities per departure or
number of fatalities per 100,000 hours flown, while automobile
fatality rates are expressed in number of fatalities per 100 million
miles driven. Using an average speed of 33.33 mph, automobile fatality
rates can be converted into hours driven.

I found an interesting risk analysis here:
<http://www.rvs.uni-bielefeld.de/publications/Incidents/DOCS/Research/Rvs/Article/probability.html>

(It even mentions Bayesian theory!).
He makes the point that some car drivers have much greater risk
factors than others, just as some airlines and some types of plane
have much greater risk factors. He doesn't seem to come to any
definite conclusions, but one suggestion was that for journeys around
600 miles, the risk was about equal for plane and car. For longer
journeys, car risk increases.


Anyway, the situation is not as clearcut as the airline industry would
like you to believe.

BillK