Car Trips vs. Airplane Flights

[barton]

I was in an agrument with a very intelligent friend about a week ago, discussing the safety of cars and airplanes. Despite my protests, he refused to believe that planes are a noticeably safer means of travel.

Now, he admitted that many more people are killed in car accidents, but he counteragrued that there are many more car trips vs. airplane flights. If those were factored into the equation, he guessed, then the wide gap between deaths wouldn't be nearly as large.

Is he right? I suspect he is, though not to the degree he was claiming. I've been looking for some statistics for this, haven't found any yet. Still, there must be over a billion car trips in the States per year - there can't be anything close to that in airplane flights.

[Whack-a-Mole]

I was JUST thinking of posting a question almost exactly like this one not two minutes ago! Some user with problems had the temerity to ask me for help. Bah! Next time the user gets his account disabled and computer re-formatted...see if he bugs me again!

On a vehicle per vehicle basis I too wonder what the percentages of accidents between planes and cars is.

However, there are a lot of ways to spin statistics if you want to determine what's safest.

1) Your chances per amount of time spent in vehicle?

2) Your chances per mile travelled?

3) Fatal vs. non-fatal accidents when they do occur?

4) Your chances as a member of the overall population (in which you'd be included even if you never flew a day in your life)?

There's probably more. My guess is the airline industry quotes stats based on #4 and maybe #2 as that would provide the best looking (for them) stats.

[bibliophage]

If you consider fatalities per 100,000,000 passenger-miles, cars are more than 20 times more dangerous than commercial airliners. See http://www.apta.com/stats/safety/natsafe.htm

TYPE OF VEHICLE
DEATH RATE

Automobiles
0.95

Intercity & commuter railroads
0.04

Airlines
0.04

Intercity buses
0.01

School buses
0.01

Transit buses
0.01

Heavy, light, & other rail vehicles Not reported

[barton]

I just thought to go to the Bureau of Transportation webpage. Given the earlier thread where, if 110 people die per day by car, and 2 by airplane, mixed with the 1995 BoT data (I like to overwhelm my friend with statistics when possible)

This is extremely rough - I'm no expert, but this is what I came to:

By raw deaths

40150 car dead per year
720 plane dead per year

cars - 98%
planes - 2%

By mileage -

827 billion miles overall in '95
451 billion by car - 57%
355 billion by airplane - 43%

57% of miles, 98% of deaths
43% of miles, 2% of deaths

by number of trips -

1 billion overall in '95
813 million by car - 81%
161 million by plane 16%

81% of trips, 98% of deaths
16% of trips, 2% of deaths


I think I win. The mileage one is pretty damning.

[barton]

I forgot to do some additional division:

57% of miles, 98% of deaths, 1 in every 11,232,876 miles
43% of miles, 2% of deaths,, 1 in every 493,055,555 miles

About 50 to 1.

81% of trips, 98% of deaths, 1 in 20249 trips
16% of trips, 2% of deaths, 1 in 223,611 trips

Here it narrows - about 11 to 1.

So, on your 20,249 car trip, walk. ;)

[RM Mentock]

Well, here's something that I bet you didn't think about.

All those per-mile statistics are pretty well established and verified--mostly because they make the airline industry look pretty good.

If you were to analyze them on a per person per hour basis though, things change. For any given individual, their chance of dying while in a car is about the same as their chance of dying while in a plane. This seems to reflect a basic human risk acceptance.

OK, you say, but for any given trip, you spend less time in the plane. Wouldn't that make the plane safer?

Not necessarily! It turns out, on the same basis, ordinary life is twice as dangerous as either car or plane. So, the quick plane trip just allows you to re-enter the more dangerous ambient environment sooner.

How is that reconciled, if the rates just reflect basic human risk acceptance? Well, the danger of ordinary life is weighted--the older or more unhealthy you are, the more risk of dying you have. You're less apt to be in a car or plane if you are old or unhealthy--you're probably going to be in a bed at home or in the hospital. Twice as less likely, it seems.

[scr4]

Not necessarily! It turns out, on the same basis, ordinary life is twice as dangerous as either car or plane. So, the quick plane trip just allows you to re-enter the more dangerous ambient environment sooner.

But the statistics given above shows that riding a bus is 100 times safer than a car. So being in a bus is 200 times safer than ordinary life! Wow.

Seriously, the accident statistics only include, well, accidents. I'm sure it doesn't include people having heart attacks in cars, etc.

[Balance]

Frankly, I don't give an attempted aerial coition with a continually axially reoriented and linearly displaced toroidal pastry about the statistics. I hate to fly. I've driven over a third of a million miles personally, and never had a serious accident--or even a particularly close call. I think I'll stick with that.

On the subject of the statistics, though, how do they come up with the "miles driven" information? I know the error in it probably isn't statistically significant, but I'm curious as to how they know how many miles we've driven.

[RM Mentock]

Originally posted by Balance
On the subject of the statistics, though, how do they come up with the "miles driven" information? I know the error in it probably isn't statistically significant, but I'm curious as to how they know how many miles we've driven.

They don't. It's a statistic.

[Whack-a-Mole]

On the subject of the statistics, though, how do they come up with the "miles driven" information? I know the error in it probably isn't statistically significant, but I'm curious as to how they know how many miles we've driven.

As RM mentioned they don't know how many miles you've driven but they can come up with statistically meaningful averages.

Just a WAG but one way might be to look at used cars. Data is easily collected on used cars -- how many miles are on the odometer and how old the car is. I'm no statistician but I imagine the sample from this alone could be considered statistically significant giving a fair approximation of how many miles are driven yearly.

[RM Mentock]

Originally posted by scr4
[QUOTE]But the statistics given above shows that riding a bus is 100 times safer than a car. So being in a bus is 200 times safer than ordinary life! Wow.

Seriously, the accident statistics only include, well, accidents. I'm sure it doesn't include people having heart attacks in cars, etc.

Exactly.

OTOH, for many people, being in a bus is ordinary life. Sick people take buses all the time. At least, in my neighborhood.

[BobT]

Even the most mismanaged airline probably maintains its planes better than the AVERAGE person maintains his/her vehicle. (Yes, some people keep their cars in perfect condition, but a lot of people wouldn't notice that their car has bald tires until they spin off a wet road and into a ditch.)

However, I believe that the majority of aviation accidents are credited to pilot error. I would assert that the average person with a pilot's license is much more skilled and better trained at handling his aircraft than you or I is at driving our cars.

Life is a crapshoot. Eventually, something is going to end our life. What control we have over that depends upon what risks we are willing to accept.

[barton]

[QUOTE]Originally posted by RM Mentock
[B]Well, here's something that I bet you didn't think about.

All those per-mile statistics are pretty well established and verified--mostly because they make the airline industry look pretty good.

If you were to analyze them on a per person per hour basis though, things change. For any given individual, their chance of dying while in a car is about the same as their chance of dying while in a plane. This seems to reflect a basic human risk acceptance.</b>

Interesting opinion.

I'll note I was looking at it from the perspective of 'what are the chances once you've decided to take a trip in this given vehicle'.

The immense mass of car-related deaths would seem to invalidate 'is about the same'. Even the most conservative measurement I've run into - the one I calculated this morning, in fact - is 11 to 1 car before plane, and it wasn't per-mile, it was per-trip.

Sure, if you'd never gotten inside a car or a plane in your life, then your chance of death by travelling in these vehicles is zero, of course. But very few people have, and you can't prove they never got into one until they've already died. The statistics reflect the people that chose to take trips.

Not sure if I understood your reasoning correctly, RM.

[tcburnett]

Originally posted by Balance
Frankly, I don't give an attempted aerial coition with a continually axially reoriented and linearly displaced toroidal pastry about the statistics.

What a poetic and mellifluous way to say "I don't give a flying fuck at a rolling doughnut about the statistics!"

[Sam Stone]

Just to be clear, the stats for airline accidents only include the major carriers. Commuter aircraft have accident rates that are much worse, almost on a par with autos. And general aviation (you in your Cessna) has an accident rate worse than autos, but not horribly so. About on par with motorcycles.

As for the major airlines, there is no way you can massage the statistics to make them look more dangerous than driving your car. You can come up with stats per trip, per mile, per hour, per takeoff and landing, whatever you want. They ALL come out better than driving a car. There's just no way to argue around it.

[tcburnett]

Originally posted by barton
I was in an agrument with a very intelligent friend about a week ago, discussing the safety of cars and airplanes. Despite my protests, he refused to believe that planes are a noticeably safer means of travel.


The numbers say airplanes are safer....at least until the Concorde went down...but no one factors military crashes into the equation. The stats refer to 'passenger' miles, which means 'paying passengers on commercial airlines'.

Even so, it is natural to think that cars are safer. When you are driving a car, you have the illusion of being in control of your fate. And you can still scream directions to the driver from the back seat if you happen to be a passenger.

An airplane is different. You crowd into a little seat and, after a suitable period of waiting to enable your adrenaline to pump up, this huge metal cylinder, wings flapping and hydraulic pumps groaning, abruptly lifts into the sky. Then the flaps and slats come up. Then the gear comes up.

A flight attendant who looks like he (or she) needs another couple Valium start passing out snacks. By this time many people have concluded that their last meal is going to be Planter's Dry Roasted Peanuts and Sprite.

The chief attendant tells you what to do in 'an emergency'. Like the plane is just going to glide slowly to a calm spot in the middle of the ocean, land safely on the water and wait while you get your floation device out from under the seat; proceed to the nearest exit and slide down the escape ramps. All the time you are thinking "bullshit, it's 4 minutes of firey hell from this altitude until I hit the concrete."

And you don't know who is driving this thing. Sure, you saw a couple guys with black uniform coats with fold stripes on the sleeves and hats go into the cockpit, but one or both may have a heart attack in 6 minutes. Or be suicidal. Od decide he needs slaves to work for him in the netherworld.

Halfway through the trip right after you gnawed at something called 'icelandic sole with figberry torte', you hear "This is the Captain...We might have a little turbulence up ahead, so if you would please remain in your seats with your seatbelts fastened and snug, the flight attendants will be passing out barf bags to those who need more than one, and some smelling salts.

I can go on for another hour, but maybe MPSIMS is a better place.

Point is, you don't feel you have control of your own life in an airplane. That's why people don't like to fly.

[RM Mentock]

Originally posted by barton
The immense mass of car-related deaths would seem to invalidate 'is about the same'. Even the most conservative measurement I've run into - the one I calculated this morning, in fact - is 11 to 1 car before plane, and it wasn't per-mile, it was per-trip.

<snip>

Not sure if I understood your reasoning correctly, RM.


That 11 to 1 ratio is pretty close to the one I calculated to produce my statistics. Look at the numbers that bibliophage posted, .95 per 100M passenger miles for autos, and .04 for airlines. Actually, the airline ratio fluctuates a great deal more than the auto ratio...some years it's three or four times that. Using 1992 info, for instance, the airline figure is .09. That ratio of .95 to .09 is approximately 11 to 1, as you have. Let's start with that.

Now, how do we calculate a statistic ratio based upon per person per hour? We divide each figure by the average speed. Since airline speed is typically ten times the auto speed, we get 1.1 versus 1.0. About the same.

In other words, your chances of dying in a car are about the same as your chances in an airline, hour for hour. On a trip basis, since auto travel takes ten times longer for the same distance, the ratio is 11 to 1, as you've said--but the overall risk of death is twice that! So, if you spend twenty hours in a car getting somewhere, you're actually better off, because after your two hour plane trip, you spent 18 hours in a more dangerous environment: life. :)

[jamshid]

I hate to fly and it shows.

Statistics have shown that flying is safer. Whoopie do da day! I'd rather take my chances on the ground, where survivability in the event of a crash is quite a bit greater--if your plane crashes, there's a percentage chance of surviving that is near enough to zero as to be virtually indistinguishable from zero.

[RM Mentock]

Originally posted by RM Mentock
On a trip basis, since auto travel takes ten times longer for the same distance, the ratio is 11 to 1, as you've said--but the overall risk of death is twice that!

That didn't quite come out the way I meant it. The overall risk of death is for all life activities...in other words, 24 hours a day. It's about twice the per hour risk for airline travel or auto travel.

[scr4]

Originally posted by tcburnett
Point is, you don't feel you have control of your own life in an airplane. That's why people don't like to fly.[/B]

Exactly. And no amount of statistics is going to convince those people, because most people think they are better than average drivers. Or these days, most people think their car is safer than average. In any case, most of us think car accident stats don't apply to us.

[Luckie]

I've heard similar statistics about commercial air travel being much safer than driving on a per mile basis, and i certainly believe the statistics. After all, there is a lot of empty air up there; in a car you whiz by trees and other cars.
But i'm not sure that comparing air travel to normal average driving is truely the most appropriate comparison. If I am comparing the danger in driving from Wash DC to , oh, say... Baton Rouge, versus the danger in taking Delta for the same trip, i don't care about the statistics for normal average driving.
Although i have no information to back it up, it seems like driving on the interstates between cities is much safer on a per mile basis than your average driving. If i can personally assure myself that i will be driving perfectly sober, during the day, and stopping for the night when i get even slightly drowsy, then my odds of fatality per mile have dropped to MUCH MUCH lower than the ones normally attributed to driving.
Now THOSE are the stats i would like to see to make the comparison. It is not obvious that commercial air travel would be safer then (though it would still be quite believable, just not obvious without doing the research).
just my two cents....
-Luckie

---------------------------------------------------------
All this and he makes music too !!??!!
sure, give it a listen
http://www.mp3.com/InSyzygy
http://www.mp3.com/RobRyland

[matt]

This article from New Scientist states that on a deaths-per-trip basis, airplanes are 12 times MORE dangerous than cars.

http://www.newscientist.com/ns/19990807/flightinto.html

[RM Mentock]

Originally posted by matt
This article from New Scientist states that on a deaths-per-trip basis, airplanes are 12 times MORE dangerous than cars.

http://www.newscientist.com/ns/19990807/flightinto.html

Cute.

What that article does, using essentially the same data, is compute the rates on a per trip basis, but not equivalent trips. Suppose, on average, we have for every one plane trip of 2000 miles, 100 car trips of 20 miles--which might not be a bad approximation, right? Then the number of miles traveled is the same, so the deaths on the car trips will be 11 times the deaths on the plane trips. But if you then compute the statistic on a per trip basis, you'd divide 11 by 100--so the ratios get reversed, and airlines have ten (OK, nine) times the deaths of automobiles on that basis.

However, he gets hung up in all this juggling. He makes the statement: "If you flew a Boeing 737 to work every morning and back, you'd be far more likely to die in a crash than if you drove a car." Now, he's clearly comparing equivalent trips. On that basis, you're ten times more likely to die in a car, because you spend ten times more time in the car making the trip. In other words, he's got it flat out backwards, at least as far as that statement goes.

[CandyMan]

I may be making a poor assumption here, but quite a few of the reasons you might die in a non-car or non-airline related accident would also apply to your time in a car or airline as well. So it can not really be twice as dangerous.

For example, a man dies of a heart attack on a plane to denver, his statistic doesn't fall under airline deaths. A guy in a passing car could shoot your buddy in the back seat.

This is the exact opposite of other statistics. Alcohol related accidents include accidents were a driver was sober, but the pedistrian he hit was drunk.

I am sticking with the theory that birth is still the leading cause of death.

CandyMan

[RM Mentock]

Originally posted by CandyMan
Life Twice as Deadly?

Yep, you got me there. You're not going to be able to separate the two completely. Life is 24/7, right? For most of us, anyway.

[CurtC]

So what we really should be comparing is whether an airplane ride is safer than driving the same trip in a car. For example, next week I'm travelling to a city that's about 200 miles away. I still haven't decided whether I would rather fly or drive. It takes slightly longer to drive this than to fly (because if I fly, I have to go to the airport, park, wait, rent a car, etc.), but an advantage of driving is that I can do it when I'm ready and not have to wait for Southwest Airlines' schedule, so I could likely get home sooner by driving.

But suppose I'd also like to compare relative risks. The statistics for the plane trip (deaths per passenger-mile) are probably skewed too low in this case, since the trip is shorter than average, and most accidents happen at takeoff or landing. If the 0.04 number is for trips that average 1000 miles, then maybe this trip should be adjusted to around 0.15.

And the statistics for car risk are probably skewed too high, since they include mostly in-town driving, which is (I think) more dangerous. The trip I will be taking is straight down an interstate highway, so it's probably much safer on a passenger-mile basis than the DOT statistics. Does anyone have any idea about the "deaths per 100 M passenger-miles" number that would apply to interstate driving? My WAG is that the 0.95 number should be adjusted down to around 0.3.

Then you'd need to figure that if I fly, I will still be doing some driving on city freeways to and from airports, and some of it will be in an unfamiliar car. All together, it seems that flying is marginally safer than driving, so I'll probably make the decision based on convenience vs. cost.