28 years after the Hurt Report, American motorcyclists are finally getting another (allegedly) major study of motorcycle accident and injury causation. Before we get too enthused, the funding is a meager $2.8M from the Fed coupled with private and state money of less than $700K to date. Worse, the Federal Highway Administration's October 5th press release states that the study will collect data on "hundreds" of motorcycle accidents. Hundreds?! By DOT's own numbers, in 1998 alone there were about 5,300 fatalities and 96,000 non-fatal injuries, so its reasonable to assume there were over 100,000 reported crashes. for that year. I believe our common experience informs us that there had to have been at least twice that number of crashes in which the rider either wasn't injured or had minor injuries for which professional treatment wasn't sought, or which weren't reported to public agencies for various reasons. I only vaguely understand the concept of statistical sampling, but is it scientifically valid to extrapolate the data of hundreds of crashes in this instance? (If you're a statistician, please comment.) My fear is that (by no fault of the researchers) we're going to see an under-funded study of a statistically too-small sample, long on correlation and short on causation, which history tells us will then be misinterpreted and/or misused by lawmakers and insurers. I hope I'm proven wrong.