Bad Statistics

The use of fishy of statistics represents a frequent an item of (dis) interest on this blog. Competitor Magazine becomes our first entrant in Dumb Use of Statistics with its 5 Myths of Bike Commuting (HT: Commute By Bike). The first myth, the roads are unsafe for cycling, contains enough hokum to be nonsensical. Let’s take a look.

You’ve got a better chance of being killed by stabbing (3 to 1) than you do of being killed on your bicycle. Ladies, saddle up: If you’re a female, you are seven times less likely to be killed on your bike than men.

How many people are stabbed? Stabbing do not simply occur. People are not randomly stabbed. A stabbing occurs within the context of a greater crime such as a robbery or a dice game. Police will classify some of these crimes as a robbery and others as a stabbing. The use of stabbing statistics is questionable.

Does the number of people killed while biking includes children? The author would have better off stating the annual number of people killed while riding a bicycle.

The inclusion of a ratio always signals a problem with logic.

Morning commutes are the safest, with the peak hours of 6 to 9am accounting for just 8 percent of all bicycle fatalities nationwide in 2005.

The lies of statistics continue. A biking population contains 10,000 people. Of those 10,000 people, 100 will die in a year while riding their bicycle. Of the 100 who died, 8 were riding their bikes between and 6 and 9 a.m. How many of the 10,000 were riding their bicycles between 6 and 9 a.m.? Eight? 10? 100? How many bicyclists were on the road between 6 and 9 a.m. to start with?

If 8% died while riding their bicycles between 6 and 9 a.m., then, by percent, a lot less died while riding between 3 and 6 a.m. The percentage between 9 and noon probably is not much higher than 8%.

Without benefit of seeing the source material for these statistics, I wager deadliest time period is 3 to 6 p.m. Unfortunately, children are not the best or smartest riders in the world.

If you wear a helmet, good for you—you’re only 10 percent as likely to be killed as those who ride without a lid.

Huh? This statement makes absolutely no sense. If a 100 people died while riding a bicycle and were not wearing a helmet at time of death, then how many people died while riding a bicycle and were wearing a helmet at time of death? 110? 90? 10?

The government’s report on the benefits of helmet use is a bit murky.

All American studies of helmet use including a report from the American College of Surgeons is flawed because they do not account for or control for many variables. How was the person riding before the accident occurred? Was traffic heavy or light? Was the driver impaired (i.e., drunk) or otherwise not paying attention (i.e., talking on cell phone or changing the radio station)? What was the rate of speed for the car at point of impact? Did the accident happen at an intersection or in the flow of traffic? These questions are not included in helmet use in studies. Helmet use studies too frequently rely on self-reports, which are problematic, or treat all bicycle accidents as the same.

Since New Zealand adopted compulsory helmet use for bicyclists, helmet use did increase but head injury rates remained unchanged.

Why the emphasis on helmet use for bicyclists? Why is no one challenging the lack of good, scientifically collected data? Bell Helmets, makers of both bicycle and motorcycle helments, pushes its agenda through public pressure groups like Safe Kids USA.

Contrast these stats to driving your car, where you’re about 40 times more likely to die than on your bike.

Start with the population for these two activities. How many people drive a car? How many people ride a bike? In America, people who ride bicycles generally drive cars at some point. However, most people who drive cars do not ride bicycles. Since the vast majority of Americans drive cars and few Americans ride bicycles, you are 40 times more likely to die in an accident while driving than while riding.

Myth: Busted.

Shit: Shoveled.