If unbiased, and proper and consistent data collection and criteria used, statistics can be very useful.
some time ago I came across the statistical report regarding the fatal skydiving accidents in US. I cannot find it now, but what I can recall was:
Chance of fatal accident per 1000 particpants:
Motorbike driving 0,2% per year
Skydiving 1,1% per year
IMHO, one of the flaws of that report was the fact that they've calculated all registered motorbike drivers and not only the active population of them.