"John_Kane" wrote in message
oups.com...
Can I get a cite on the Utah study ? It looks interesting


The relevant data is from Table 51 of NHTSA's State Data System Crash Data
Report: 1990-1999, which can be found at
http://www-nrd.nhtsa.dot.gov/pdf/nrd.../15safetyequip
..pdf

Utah is one of 17 states contributing data to the SDS, which is a database
similar to FARS, except that it includes *all* injury-related traffic
crashes, not just those resulting in a fatality. There is a considerable
problem with the Utah MC helmet data, in that for about 72% of both the
fatal and non-fatal cases from 1990-99, helmet use status is listed as
"unknown." Nonetheless, if the remaining 28% of the cases where helmet use
status is known are actually representative, we have 43 helmeted fatalities,
5 non-helmeted fatalities, 1561 helmeted non-fatalities, and 185
non-helmeted non-fatalities, which yields an apparent odds ratio of
(43/5)*(185/1561) = 1.02 -- or essentially a null effect (i.e. helmeted
motorcyclists being about equally likely to suffer a fatal injury relative
to non-helmeted motorcyclists). OTOH, because of the low number of
fatalities involved, the result is exceedingly unstable. For example,
changing just a single "Helmet use unknown" fatality to a "Helmet not used"
fatality reduces the apparent odds ratio to 0.85 (a 15% net safety benefit),
so these results should be treated with considerable caution.

Riley Geary