33

Considering the number of people that use it compared to other transports (cars, trains, boats), do you have less chances to die in a plane?

Are there any studies about this?

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
Nikko
  • 2,975
  • 2
  • 20
  • 17
  • 12
    When I've looked into this in the past it was necessary to be precise about what you mean by "safest". Supporters of this claim like to examine "per mile" statistics, and detractors "per trip" stats, and even then you need to consider regional variation across the globe. – dmckee --- ex-moderator kitten Aug 03 '11 at 22:07
  • 1
    You're right, all these variations are intersting. Maybe it would also be of interest to compare deaths and injuries together and separately for each mode. – Nikko Aug 04 '11 at 07:59
  • 1
    Simple fact that's totally over looked here is death per trip. If you get in a car and drive, you are 4 times less likely to die(if you are the average driver). If you are a very good driver, you are most likely upto 1000 times less likely to die. –  Apr 15 '15 at 00:33
  • @EdwardDNewgate: If you want to compare driving between two distant cities and flying, "death per trip" is misleading. Trips tend to be longer (distance and time) in planes. "Death per million kms travelled" is more accurate for long trips. See the accepted answer for more details. – Oddthinking Apr 15 '15 at 05:13
  • Not "airplanes", but "airlines". General aviation is still quite dangerous. – vartec Apr 15 '15 at 20:21
  • The answers below don't address whether spaceship travel is safer than air travel. – Andrew Grimm May 18 '19 at 01:47

2 Answers2

43

The question of airplane safety is a nice field to demonstrate the old quip that you shouldn't trust any statistic you haven't forged yourself.

It is crucial to define exactly what you mean by safe, and what you are comparing:

Are you interested in total deaths per year? Then the car is the most dangerous and, depending on the country you are looking at, trains are the least dangerous, with airplanes also quite safe:

http://www.medicine.ox.ac.uk/bandolier/booth/Risk/trasnsportpop.html

Your annual odds to die in a passenger airline accident is 1 in 4,406,209 in the US, whereas the annaul odds to die in a car accident are 1 in 6,478. This seems to indicate that a plane is almost a thousand times safer than a car.

But is that the question one should be asking? You could ask: What is the chance of death in a transport related accident per hour of travel, or per kilometer traveled.

I found this short statement in a review on the book "Freakonomics":

Levitt also debunks some of today's more relevant conventional wisdom: the notion of driving being a far more dangerous form of transportation than flying. While it's true that many more people die annually in car accidents than in plane crashes, it's often overlooked that the dramatic difference in number of deaths is largely due to the amount of time the average person spends in an automobile in comparison to the relatively small number of hours spent in flight. Levitt goes on to show the per hour death rate of driving to be about equal to that of flying. Have a nice trip.

So what is the conclusion we can draw? Well, you can convert the chance of death per hour into a chance of death per trip by multiplying with the duration of the trip. And even if both car and plane have the same hourly rate, a trip in a plane for the same destination is far shorter than a trip in the car. If the hourly death rates are the same, then a plane is safer on a per trip basis by how much faster it is than a car. While this still means that a plane is safer than a car per trip, the factor is far less than 1000. More on the order of ~10.

Lagerbaer
  • 12,311
  • 2
  • 68
  • 81
  • Great answer Lagerbaer. Not much I could add to it. – Larian LeQuella Aug 04 '11 at 02:13
  • 16
    The problem I see with these kinds of statistics, is they deal with averages. In case of car travel there is really high correlation between risk and risky behaviors. For example some 25% of fatal accidents are caused by 5% youngest drivers. – vartec Aug 04 '11 at 09:29
  • I don't accept this answer right away, I let some time pass if someone wants to add another answer. – Nikko Aug 04 '11 at 11:45
  • 1
    @vartec brings up a good point: [Thrill seekers, nonconformists, and people with anger issues are more likely to get in car accidents](http://www.sciencedirect.com/science/article/pii/S0191886902000107). – ESultanik Aug 04 '11 at 12:01
  • and hopefully they do not pilot airplanes? Maybe it's hard to compare a transport mode using professional drivers with ordinary people driving cars – Nikko Aug 04 '11 at 12:19
  • 2
    Do you have comparable statistics for bicycling? – Apreche Aug 04 '11 at 12:20
  • 3
    "What is the chance of death in a transport related accident per hour of travel, or per kilometer traveled." I would ask exaclty those two questions (per hour and per km). Would you happen to know the answer? – Suma Aug 04 '11 at 13:07
  • 1
    @Suma I tried some time googling but did not find exactly what I was looking for, especially not for cars. I'll try harder :) – Lagerbaer Aug 04 '11 at 14:44
  • @vartec You are absolutely right, and it would make sense to further split up the "car accidents" statistics into gender and age. – Lagerbaer Aug 04 '11 at 14:45
  • Interesting would be the risk of dying by a traffic accident (car/train/airplane) for people which did not sit *in* the car/train/airplane, but were simply "collateral damage". I'm not sure how to quantify this. – Paŭlo Ebermann Aug 05 '11 at 00:44
  • @Paŭlo Ebermann For cars, this would still count as a traffic related accident and should show up in statistics. For planes, this would mean that the plane crashed on you. I believe that this number is negligible and thus you could recycle the "normal" plane numbers – Lagerbaer Aug 05 '11 at 02:05
  • So driving by car might actually be safer (for the driver) than statistics pretend? – Paŭlo Ebermann Aug 05 '11 at 02:08
  • Depends if the statistics count you getting run over by a car as a death while driving a car, which would be inaccurate. The link I give does it right, I think, since it has its own category "Pedestrian killed in road traffic accident". – Lagerbaer Aug 05 '11 at 03:49
  • 1
    @Lagerbaer: A car accident can affect pedestrians or bikers as well. The question of own risk is important (think: alcohol) and the problem, that most short distance travelling can't be done by airplane. This wouldn't be important, if the risk could be calculated, because it was independent of the distance, but I guess it is not independent. AFAIK, highway travelling is more secure than in the cities or on smaller streets. And risky behaviour isn't independent from distance (driving home from the pub or disco, drunken, is almost below 100 km). – user unknown Aug 05 '11 at 08:47
  • 2
    ... and: Since airports aren't the real targets, there has to be done some calculation, because not every kilometer flown was in the interest of the flying person. From start to stop, you normally have a distance start -> Airport A -> Airport B -> stop. This could be compared and counted as the distance start -> stop, as if it was done by car. – user unknown Aug 05 '11 at 08:53
  • @user unknown I don't think so. Remember the whole point of these statistics is mostly to help people overcome their fear of planes. I don't see how car trips to and from the airport would be relevant to that. – djeidot Aug 07 '11 at 12:20
  • If your point is, to overcome the fear of somebody, independent of all logic, you lost me. But since planes aren't bound to streets, they might as well use shorter ways than alternative transport forms. – user unknown Aug 07 '11 at 13:26
  • I really don't trust these statistics. 1 in 6478 people is 1,157,000 car crash deaths per year. – tuskiomi Jul 12 '17 at 18:37
12

The chance per trip stats are more complex than suggested in other answers suggest. The normal car accident stats are for average car trips, so are the air stats. Thus the normal car stats are for short average trips to and from work and air for long average trips between countries and over seas.

From Wikipedia http://en.wikipedia.org/wiki/Air_safety#Statistics

It is necessary to mention that first two statistics ["Deaths per billion passenger-journeys and "Deaths per billion passenger-hours"] are computed for typical travels for respective forms of transport, so they cannot be used directly to compare risks related to different forms of transport in a particular travel "from A to B". For example: according to statistics, a typical flight from Los Angeles to New York will carry a larger risk factor than a typical car travel from home to office. But a car travel from Los Angeles to New York would not be typical. It would be as large as several dozens of typical car travels, and associated risk will be larger as well.

Flying from London to Beijing would be quite normal for a plane but highly unusual for a car.

Such a trip in a car would be fraught with dangers not normally shown in the average car stats such as tiredness, extreme wear and tear on the car, and even highly unusual things such as car hijackers, thieves, bandits and war zones. Travel by air avoids almost all of these hazards.

Conversely, the most dangerous times when flying are in the approach/landing (9%/38%) and taking off (14%). Short, car like, trips would increase the amount of landing and taking off per mile travelled away from the average stats for the plane. Flying to work and to the shops, instead of driving, would probably be relatively dangerous compared to driving. This probably being of of the key reasons that flying cars have never and will never become popular even if technically possible.

Rincewind42
  • 852
  • 5
  • 18
  • This isn't an answer. It is a comment. Please make a comment and delete this. – Oddthinking Aug 05 '11 at 01:25
  • 6
    Added references as asked. This isn't a comment, it's an answer. Air planes are by far the safest for intercontinental travel by miles and miles. However, Air planes are not a safe way to travel to the shops for groceries. – Rincewind42 Aug 05 '11 at 02:00
  • 1
    So ideally, you should take the plane for long distances, and just use the car instead of taking off and landing ;) – Benjol Aug 07 '13 at 11:23