Wavepacket Blog displaying only one specific post |
| 2008 | |
| June | |
| Sun Jun 8 16:39:28 2008 Forbes: Fooled by Randomness! |
|
| >> links >> | |
| Sun Jun 8 16:39:28 2008 Forbes: Fooled by Randomness! Forbes flubs basic statistics. |
|
| In my previous entry (
Fooled By Randomness) I reviewed a book that pointed out all of the mistakes people make with
statistics.
I had to laugh, then, when Forbes came out with their list of most dangerous sports (see the slideshow here). The list is NOT the list of the most dangerous sports! It is just a list of where the most sports injuries occur. As the authors point out: "The rankings do not take into account varying participation rates, which partly explains why many popular sports yielded the greatest number of injuries." Put another way: the Forbes list gives you NO idea which sport is the most dangerous. A more accurate name would be "The most popular sports in the US as measured by injuries." For instance, on their list, jumping out of a plane without a parachute would count as a very safe sport, since it happens so rarely. And BASE Jumping didn't even make the list. Basketball (!) was what Forbes considered the most dangerous sport, well ahead of American Football and BASE Jumping. The list is interesting, but clearly it doesn't tell you anything about danger. (I mean, come ON! Watch this BASE jumping video and tell me that basketball is remotely dangerous.) In order to determine how dangerous a sport is, you need to figure out NOT the absolute count of injuries, but the RATE of injuries. A better question is: "If I play this sport regularly for a month, what is my chance of getting injured?" (Where "regularly" depends on the sport). If you do searches for Dangerous Sports, you'll get a bunch of useless results. Again, people confuse the absolute number of injuries with the rate of injuries. I started to get worried. Maybe no one who studies sports has any concept of probabilities? For a second, I had high hopes for a 2003 article that studied sports injuries in the US from 1997 - 1999 (you can find the article at The British Medical Journal's website). It claimed to have normalized data, with injury rates. However, all they did was divide the injury rates by the population size! So again, popular sports ranked highest. Not surprisingly, Basketball was at the top of the list. That was sobering. Yikes! Even doctors studying sports injuries had gotten it wrong. (Although, to be fair, they weren't trying to compile a list of the most dangerous sports. But wouldn't that be useful?). More digging turned up a few numbers. This book had some useful data, although limited. It noted that the most dangerous sports it surveyed were Rugby (59.3 injuries per 100 participants in 4 weeks), Soccer (39.3), Martial Arts (36.3), Hockey (24.8), and Cricket (20.2). That is exactly the sort of data that can help you define dangerous sports! And I don't think it's too much surprise that rugby is near the top. (Obviously that was a British book). Then this link had data on skiing and snowboarding. They measured rates in terms of injuries per 1000 ski days (on average, how many people are injured on a day with 1000 people on the slopes, for instance). To convert that to the other metric (number of injuries per 100 participants in 4 weeks) I had to guess how often your average skier/snowboarder went per week. Assuming 10 visits to the slopes in 4 weeks (all weekends and two extra days), that yields an injury rate of 3 injuries per 100 participants in 4 weeks for Skiing, and 4 for snowboarding. (I think those numbers are slightly lower because the study only counted injuries requiring medical attention). Based on that site, Skiboarding had rates of 8 injuries per 100 participants in 4 weeks, while Nordic Skiing was lowest around 0.4. Then this article had more data. According to the NCAA, the most dangerous college sports are spring football (9.6 injuries per 1000 participants), men's wrestling (5.7), women's soccer (5.2), and men's soccer (4.3). Men's fall football was only 3.8 injuries per 1000 participants. To convert those numbers (injuries per 1000 participants) to the above (injuries per 100 participants over 4 weeks) I had to guess how many weeks people participated in college sports. I assumed 12 weeks per season. Doing the math (converting 12-week injury rates to 4-week injury rates) yields 0.32 injuries per 100 participants in 4 weeks for spring football, 0.19 for men's wrestling, 0.17 for women's soccer, 0.14 for men's soccer, and 0.13 for men's fall football. I searched around a bit more, but was unable to come up with much more data. Here then are my results. Most Dangerous Sports This is an Internet exclusive! I don't think anyone has a decent compiled list of actual dangerous sports (again, other lists are just most common injuries due to popular sports). The most dangerous sports, measured by the number of injuries per 100 participants over 4 weeks:
That's not a very good list, because so many popular sports are missing, and I had to make a lot of assumptions to put it together. But that's because no one has measured or compiled the data in any useful form. Despite years of Forbes lists, it turns out that no one, not even sports physicians, have actually determined what the most dangerous sports actually are! Comments |
Related: mathematics Unrelated: books economics energy environment geopolitics lists predictions science |
| Links: |  |
Blog Directory | Blog Blog | Technorati Profile | Strange Attractor |