The Effect of Bicycle Helmet Legislation on
Bicycling Fatalities - Grant and Rutner.

Their statistics are sound, and their calculation of a 15% reduction in
the juvenile bicycling fatality rate during the helmet-law era appears
to be accurate, although virtually indistinguishable from the
already-existing downward trend since 1975, represented by the blue line
in their data graph:

http://img22.imageshack.us/img22/9715/graph.gif

They conclude that their calculated reduction is not due to any
corresponding reduction in bicycling among juveniles, but their opinion
is unsupported by the data. The indirect evidence they infer from a lack
of increase in vehicle-miles per capita, and in the juvenile pedestrian
fatality rate, is insufficient to offset the direct evidence easily
inferred from the significant increase in the rate of juvenile obesity
from 11.3% [1988-1994] to 15.3% [1999-2000]:

http://www.cdc.gov/nchs/data/hus/tab...3/03hus069.pdf