Lately this ridiculous blog post about cycling without a helmet has been making the rounds on the internet. And it’s got everything necessary to stir people up: statistics, a contentious subject, and one of those “this is so counterintuitive that it just has to be true” kind of conclusions.

It is also wrong. Wrong, wrong, wrong. 100% wrong. Wrong on the facts, wrong on the analysis, and wrong on the logic. What’s so wrong with it? Let’s go point by point.

First, the author writes, “if you get into a serious accident, wearing a helmet will probably save your life. According to a 1989 study in the New England Journal of Medicine, riders with helmets had an 85% reduction in their risk of head injury and an 88% reduction in their risk of brain injury.”

He then turns around and immediately discounts the overwhelming evidence that helmets protect people from serious head and brain injuries. He argues that because helmets only protect you if you actually have an accident — and because helmets might also be beneficial in other situations where they are not commonly worn — we should ignore this fact.

But this is exactly the wrong conclusion, the correct one is the obvious one, the one he states at the very outset of his piece: if you are in an accident, the evidence shows that wearing a helmet greatly improves your chances of surviving the accident without a debilitating injury.

Taken this way, the benefits of wearing a helmet are clear. (Let’s discount the aerodynamic benefits of wearing a helmet, since this is really a piece about casual cyclists who don’t care too much about saving 5 watts.) Now we can turn to the question of the risks and costs of wearing a helmet and see if the benefits outweigh the risks.

Here’s where the analysis goes completely off the rails — as if throwing out the 85% reduction in head injuries associated with helmet use wasn’t bad enough already.

He turns to a 1979 study of head injuries to show that most head injuries result from motor vehicle accidents. His analysis of the conclusions in this study is bad for two reasons. First, without normalizing the number of head injuries by the number of trips taken or distance traveled, he fundamentally fails to accurately assess the real risk of each activity¹.

To see just how silly this is, consider an extreme example. The number of people who die on Mt. Everest is vanishingly small. Compared to the numbers of people who die in cars, on bikes, in wars, violent crime, or because of disease, basically nobody dies on Mt. Everest. So by the author’s logic, we should conclude that climbing Mt. Everest isn’t dangerous, right?

No. We should not. The death rate on Mt. Everest is worse than 1 in 20. It is extraordinarily high. Climbing Mt. Everest is one of the most dangerous things you could possibly do. But if you don’t normalize the number of deaths by the number ascents, the you cannot compare the number to other human pursuits.

Likewise, without any normalization, we have no sense of the real risk of head injuries for cyclists versus drivers. But many more people travel by car than by bike (probably even more so in 1979) while the reported share of injuries to cyclists in the 1979 study is more than 10% of that for people in cars. This suggests that, properly normalized, the risk of head injury per trip for cyclists would be much, much higher than that for occupants of motor vehicles.

Which brings us to the other mistake he makes in his analysis of this study: in 1979 traveling in a car was really dangerous. So dangerous, in fact, that automakers dedicated a huge amount of resources to developing safety technology. Since 1979 we’ve seen the advent of better seatbelts, better child seats, airbags, crumple zones, ABS brakes, electronic stability control, the list goes on and on and on.

As a result, since 1979 the normalized rates (both per km traveled and per total population) of motor vehicle deaths have plummeted, dropping by close to 70%. Which means that a more modern study would likely show that the share of total injuries to cyclists compared to motor vehicle occupants is even worse today.

He goes on to cite evidence that the rate of head injuries in snow sports has increased, even as more and more people have begun wearing helmets on the snow. But as any Statistics 101 student knows, correlation is not the same thing as causation². Injuries in snow sports are likely the result of the changing nature of both equipment and the growing popularity of more dangerous styles of skiing and snowboarding. Aggressively shaped skis allow novice skiers to move more quickly on to more challenging terrain than they could have a couple of decades ago. Likewise, there were many fewer half-pipes and terrain parks two decades ago than there are now. And more and more people are skiing off piste, in terrain where the risk of injuries is high in general. None of these factors have much to do with helmet use. The argument that the rise of helmet use in snow sports has somehow driven an increase in numbers of head injuries is totally specious.

Meanwhile, the claim that helmet use may increase the danger from drivers themselves is also rather suspect. This is a frequently cited claim among anti-helmet advocates, but it is based on a single, poorly designed study involving just one cyclist. How do we know, for example, that the cyclist himself didn’t behave differently while wearing a helmet and not? Where are the experimental controls that would protect against the possibility that the statistics are contaminated by other subtle changes in behavior that have nothing to do with helmets? This study is intriguing, for sure, but hardly justification for discounting the overwhelming evidence of what helmets can do for you.

The author then makes another specious argument: that helmets may increase certain types of neck injuries in accidents. A more likely explanation is that accidents that may be catastrophic without a helmet are not as serious with one, meaning less serious injuries than skull fractures and brain injuries receive more attention.

He concludes by citing a study of mandatory helmet laws in Australia that suggested these laws may have reduced the number of people actually cycling. This is a concern, but mandatory helmet laws are not the same as simply encouraging people to wear helmets. Nor is the possibility that the existence of helmets might discourage a few people from taking of cycling a reasonable argument for consciously rejecting helmets altogether. Lots of people find seatbelts objectionable, but you would be hard-pressed to find evidence that mandatory seatbelt laws have done anything to discourage people from driving.

I could go on: about how the risks of cycling vary enormously across different cultures that do value bikes and have developed safe infrastructure for them, about the number of times a helmet probably saved my life and the one time I crashed without one³, or about how foolish it is to reject a safety device that has been shown to reduce the rate of catastrophic injury by 85 or 90%. I’ll spare you.

Of course, none of this is to say that you should never ever throw your leg over a bike without a helmet. If you want to do that, fine. It’s hardly the end of the world. But to make a conscious decision not to wear a helmet all the time, then try to justify that choice with a lot of incorrect statistical analysis and false logic is foolish. And to go further and call your rationalization of your risk taking “Why It Makes Sense to Bike Without A Helmet” is patently ridiculous.

¹ He sort of acknowledges this point in a dismissive addendum at the end of the article, but since his entire argument appears to be predicated on this point, no number of mealy-mouthed addenda can save him now.

² Really, correlation and causation are not the same thing.

³ Once was enough for me. I crashed, heavily, in a solo accident that was entirely my own fault. I have no memory of the crash, and very little memory of the hours before and after it. I was hurt badly enough that I couldn’t get home on my own. Then once someone did help me to get home, obviously seriously concussed, I got lost while I was looking for the bathroom in my own home. Needless to say, I learned my lesson.

I try to practice what I preach. My kid (photo above) is learning to ride a bike with a helmet.