I'm only going to ride streetcars from now on

May 30, 2008 9:32 AM

Inspired by this question - how does one really asses how "safe" a mode of transportation is?

For example, the National Safety Council calculates odds of dying for all sorts of modes of transportation (here) - is cycling really 18 times safer than riding in a car? Is riding around in a 3 wheel vehicle really 460 times safer than walking around on foot? Or are those numbers simply a result of the fact that relatively few people bicycle, and almost nobody runs around on 3 wheels?

posted by NoRelationToLea to travel & transportation (15 comments total)

It looks like for all of those "one year odds" numbers, they just divided the number of fatal accidents annually by the population of the United States (about 300 million). The "lifetime odds" are then just the annual odds divided by about 79 (life expectancy, I suppose). A pretty amateurish analysis, really. I don't think it remotely approaches statistical validity, and it certainly doesn't seem to correct for frequency of use of each mode of transport.
posted by mr_roboto at 9:48 AM on May 30, 2008

And now that I've put away my calculator, I see that the methodology is explicitly written down on that page. Yeah, that's no good.
posted by mr_roboto at 9:50 AM on May 30, 2008

Or are those numbers simply a result of the fact that relatively few people bicycle, and almost nobody runs around on 3 wheels?

Yes, that's almost certainly part of it. You'd have to know how many people bicycle, how many miles they bicycle, and derive your numbers from there. The table you linked is comparing apples and oranges. I poked around the Bureau of Transportation Statistics site and didn't find anything definitive, but if you're willing to do some math, you could find vehicle miles traveled, etc.
posted by desjardins at 9:54 AM on May 30, 2008

It depends on how you define "safety". You are much more likely to die driving 1000 miles than flying 1000 miles. But if you change your perspective and look at time instead of distance, you actually have pretty even chances of being killed flying for 10 hours or driving for 10 hours.

(from Freakanomics, also mentioned in this ask mefi thread).
posted by steveminutillo at 10:17 AM on May 30, 2008

I heard on CBC the other day a guy talking about this. He said that on any given day, driving is safer than cycling. But, on average over your life, cycling is far and away safer due to the health benefits from the exercise. You will live longer cycling to work instead of driving.

At least, thats the way his numbers were pointing. Sorry, can't find the reference.
posted by maxpower at 10:17 AM on May 30, 2008

The data is titled "Odds of Dying," not "Assessment of Safety."

It's perfectly accurate. It's not comparing apples to oranges. It's not "no good."

It's just that none of you guys are paying attention to what the dataset is actually measuring.
posted by gum at 10:21 AM on May 30, 2008

It's just that none of you guys are paying attention to what the dataset is actually measuring.

So what is it measuring? As best as I can tell, the "lifetime odds" column gives the probability that an randomly selected American had died in a accident in the past year, multiplied by 79. It's a pointless number.
posted by mr_roboto at 12:55 PM on May 30, 2008

It can be hard. There are a lot of different statistics. I've seen it measured per mile, per time, per trip, lifetime...

In addition, it is necessary to note that these are statistical measures and don't necessarily apply to an individual. Accident numbers include those who travel drunk (cars, bicycles, walking, being drunk increases the risks of them all) and who travel in an unsafe fashion, and those who travel in all sorts of environments. If, for example, you are not drunk, ride safely, and live on a very lightly traveled farm road your bike riding will probably be safer than average - this can go both ways.

I'm not going to go through it again because it took me forever, so take this as anecdote, but I examined myriad statistics once and the vague results I tended to get was that bicycling (helmeted or unhelmeted - I tried to get unhelmeted by assuming both types of riders experience accidents at the same rates but helmets are 100% effective - the bias from the first could go either way but the latter bias definitely made things look riskier) seemed to come out order of magnitude comparable to driving.
posted by TheOnlyCoolTim at 1:48 PM on May 30, 2008

No, it's not a pointless number. It's a way of answering the question, "What are the odds that an American will get killed in a car accident?" by replying, "Based on stuff people died from last year, one in 84." It's not the data's fault if people are asking the wrong questions.
posted by gum at 1:50 PM on May 30, 2008

how does one really asses how "safe" a mode of transportation is?

Not intending to be annoying (hopefully more like empowering) but I'd think that any difference in the inherent safety of different modes of transport is utterly dwarfed by the difference that you can make by always being alert, trained, and defensive, while commuting.
(though this can't apply to a bus or plane or train since you're not driving, but (with the exception of flight on a per minute basis) these modes are so safe they trump any mode that you can drive anyway).

So to evaluate safety, you need to look at the biggest factor in safety - you. Are you the sort of person who in a car would end up putting fuel-sipping driving techniques ahead of defensive driving techniques? Are you the sort of person who would save effort in cycling by slowing the minimal amount needed to get where you're going, thus requiring the least effort getting back lost speed? What kind of shortcuts are likely to seduce you and influence your habits and techniques? While modes of transport are most impacted by your safety-peccadillos?

(Further complications - per minute spent traveling in the vehicle: cars are safer than planes... if you count fatalities. But planes are safer than cars if you count injuries. (ie if you are in a car, there is such a thing as a "minor accident", if you're in a plane, it's much more likely to be all-or-nothing. (Along similar lines, people say that rattlesnake bites are almost never fatal. This makes them sound less dangerous - until you see the permanent wreckage that surviving entails)).
posted by -harlequin- at 2:43 PM on May 30, 2008

For comparing bicycling risk, one also needs to consider that accident and fatality statistics for the USA might include children (who might often bicycle, but rarely, if ever, drive a car) together with adults (who bicycle more rarely and drive a car much more often), making it even trickier to draw any firm conclusions.
posted by normy at 3:40 PM on May 30, 2008

TheOnlyCoolTim:
the vague results I tended to get was that bicycling... ...seemed to come out order of magnitude comparable to driving.

Safer or riskier?
posted by -harlequin- at 4:54 PM on May 30, 2008

This is a super relevant essay that breaks it down for cars vs. bicycles per-mile and per-hour: http://www.kenkifer.com/bikepages/health/risks.htm

And yeah, factor in the lifetime health benefits and bikes come out pretty far ahead.
posted by Skwirl at 6:00 PM on May 30, 2008

All over the place depending on what statistics I found and crunched.
posted by TheOnlyCoolTim at 7:04 PM on May 30, 2008

Not counting in the risk of death from obesity and poor fitness due to motor vehicle, of course.
posted by TheOnlyCoolTim at 7:07 PM on May 30, 2008