"Gary Drescher" wrote in message
. ..
Interesting data point, but it's addressing a different question. It's
looking at the accident rate per elapsed time (per year), rather than per
time spent driving. So it could be that the ones with lower accident rates
simply do less driving, rather than that they're safer when they drive.

Indeed, and it only looks at professional pilots (possibly even only airline
pilots, given the all-too-common bias against other professional pilots
among people outside the aviation industry). If it's only airline pilots,
there's a clear bias there, given that the daily routine of an airline pilot
is often VERY different from that of most of professionals, with a lot of
time spent actually in an airplane or at an airport, at least when on duty.

Of course, the article also doesn't tell us what the sample size of each
population (professional) group is. It's not hard to imagine the
possibility that they didn't even have 1000 pilots in the study, making the
statistical error of that group (and similarly under-represented groups)
much higher than for other groups.

Note also that the study was done by starting with a database of 1 million
accidents, and then cross-referencing that with a database of insurance
policy owners. This is exactly the kind of statistical analysis that others
have complained about in this thread. Personally, I think it's useful to
the extent that one recognizes its limitations, but it's not going to
"prove" anything, especially to someone insistent on ignoring the data.

Beyond that, I think it's telling that while there appear to be genuine
statistical differences, even if one assumes that they are due entirely to
individual behavior rather than circumstantial conditions, there's really
not that much difference across the various professions, especially for the
"accident" category. For an insurance company, I suppose the difference
between 80 accidents in a year for 1000 people and 100 accidents in a year
could be useful information (insurance companies live and die on aggregating
huge numbers of experiences in order to get predictable outcomes), but what
that really says is that there's a lot of overlap in individual performance.

Even if one assumes that study is completely applicable, if anything what it
suggests is that *most* pilots are probably about the same sort of driver as
*most* other individuals, at least when it comes to having accidents.

Pete