.comment-link {margin-left:.6em;} <$BlogRSDURL$>

Monday, June 19, 2006

Misunderstanding Statistics 

As Wired Cola's Director of Security, Erick, will attest, I am nearly innumerate.

As just about everybody who knows me will attest, I hardly ever ride a bicycle or motorcycle without a helmet. I have my reasons, but after reading the numbers in this article about motorcycle fatality rates in Florida, I'm beginning to wonder about how compelling the case for helmets is.

More importantly, I'm pretty sure the statistics presented in that article mean exactly the opposite of what the reporter thinks they mean.
Motorcycle fatalities involving riders without helmets have soared in the nearly six years since Gov. Jeb Bush repealed the state's mandatory helmet law, a newspaper reported Sunday.
Okay, so Florida doesn't have a mandatory helmet law any more, and now more of the dead bikers aren't wearing helmets. I already have one question...
A Florida Today analysis of federal motorcycle crash statistics found "unhelmeted" deaths in Florida rose from 22 in 1998 and 1999, the years before the helmet law repeal, to 250 in 2004, the most recent year of available data.
Hey, that sounds pretty bad! So, like, these riders are probably dying because they're not wearing helmets. I bet we could tell if you told us how this death rate compares to helmet-wearing rates in the state.
Total motorcycle deaths in the state have increased 67 percent, from 259 in 2000 to 432 in 2004, according to National Highway Traffic Safety Administration statistics.
Wow, this is just worser and worser! So deaths have risen like crazy of late. So, the level of ridership has presumably been pretty constant over that time period, thus making these shocking numbers comparable to each other, right?
Records, though, also show motorcycle registrations have increased 87 percent in Florida since Bush signed the helmet law repeal on July 1, 2000.
Oh! So what you're saying is that per-rider fatality rates have declined since the helmet law was repealed. You know, I would have thought that was the sort of counterfactual result that might have, oh, made this an interesting news item!

The rest of the article is unimportant filler.

I'd like to be clear about one thing which I have avoided saying so far: motorcycles are authentically hazardous to the health of their riders. There's lots you can do to avoid becoming a statistic, but the numbers I have seen suggest that motorcycling is about ten times more risky (per capita) than being in a car. (Cycling seems to be closer to driving than motorcycling, risk-per-capita wise, but there are some ways of fudging the measurements to make it look nearly as safe as driving. If you include the net health benefits of riding a bicycle, you might come out ahead.)

Extra credit: construct a plausible hypothesis for the decline in per-rider fatalities in Florida. I can think of two, though I don't have enough data to test either one.

So there you go. Isn't innumeracy fun?

Perhaps we need to be shown that the new registered riders are riding as much (or more) as the old riders. Otherwise, it seems like there's still too little information to draw on...
Quite right: one of my plausible explanations was that new riders were not riding as much, or were somehow riding more safely (maybe the new riders are largely fresh-off-the-jet retirees who put 20 miles per month on their Harleys).

But given the stats we do have, I would have titled this article, "despite helmet law repeal, motorcycling is less deadly."
You're right that the article is poorly written and misleading.

1. Fatalities per helmeted-rider DECREASED by 59% from 2000 to 2004.

2. Fatalities per helmetless-rider INCREASED by 508% from 2000 to 2004.

3. Overall, fatalities per rider have DECREASED by 11%.

This may be due to a) increased rider awareness of safe riding, b) increase car driver awareness due to bikes increasing usage, c) safer helmets, or maybe d) increased traffic resulting in slower riding.

Instead of "Biker fatalities soared" the headlines should read "Helmetless biker fatalities soared"
Or it may just be due to a decrease in riders riding while intoxicated...

Out4blood: you seem to have a separate source for the fatality-per-capita rate of helmetless riders in Florida; unless I'm mistaken, it can't be inferred from the numbers in the referenced articles.

If so, I'm baffled by the per-capita increase in helmetless rider fatalities in FL; this suggests that not only did more riders go helmetless (and causally or not, die) after the law was repealed, but that going helmetless in FL suddenly became vastly more dangerous than it was before the law changed.

Am I misunderstanding this?

The CDC data is interesting for a lot of reasons, especially in the footnote: I didn't realize that motorcycling had become so much more popular in the US! New motorcycle sales increased by 260% from 1997 through 2003. Assuming those sales are a reasonable proxy for the number of bikes on the road, motorcycling has become much safer, too (though by no means safe, relative to car driving): deaths per new motorcycle in 2003 were only 64% of the same figure in 1997.
Post a Comment

Links to this post:

Create a Link

This page is powered by Blogger. Isn't yours?