よく言われるのが「飛行機は一番安全な乗り物」。
それって本当たろうか?

英国王立統計学会(the Royal Statistical Society),
米国統計学会(the American Statistical Association)
Wiley-BlackwellによるウェブマガジンSignificance

の記事で乗り物の安全性を統計データで検証している。

死亡率を算定する場合の分母の取り方で意味は変わってくる。
以下の検証データは(1)乗車一回あたり (2)乗車時間あたり (3)輸送距離あたり の3通りでランキング。

単位:人/億(回/時間/距離km)

①乗車一回あたり...journeys
バイクMotorcycle: 1,640
自転車Bicycle: 170 
飛行機Air: 117 
Water: 90  
徒歩Foot: 40
車Car: 40  
トラックVan: 20  
鉄道Rail: 20
バスBus: 4.3  

②乗車時間(h)あたり ...hours
Motorcycle: 4,840人
Bicycle: 550
Foot: 220
Car: 130
Van: 60
Water: 50
Air: 30.8
Rail: 30
Bus: 11.1

③輸送距離(km)あたり...kilometres
Motorcycle: 108.9人
Foot: 54.2 
Bicycle: 44.6 
Car: 3.1 
Water: 2.6 
Van: 1.2 
Rail: 0.6 
Bus: 0.4 
Air: 0.05

と、どれを分母にするかによって3通りの死亡率が導かれる。

乗車回数を基準にすると・・

バイクが1,640人/億回でダントツで危険!!

自転車(
170人/億回)、飛行機(117人/億回)も危険(笑)

バス(4.3人/億回)が一番安全な乗り物か?


以下、原文
-------------------------------
 Sky-high? Your risks when taking a freefall
Written by Michael Wallace on 14 October 2012. Posted in Science & Technology
 
Last weekend Felix Baumgartner set a world record for the highest ever skydive, casually hopping off a skateboard-sized platform some 39 kilometres above the New Mexico desert. Millions around the world watched (a record 8 million people in tuned in to Youtube to watch Baumgartner's feat) with bated breath as the Austrian adrenaline addict tumbled through the Earth's upper atmosphere, breaking the sound barrier in the process, before deploying his parachute and coasting to an elegant landing several minutes later. Never in doubt, as they say.

Whilst the true risks of such extreme experiments are seldom calculable, the jump did make me wonder just how dangerous a (normal) skydive is. After all, if they're relatively safe I might be better off getting my own balloon-borne space capsule and simply parachuting to work rather than taking the Tube (the subway/metro to the non-Brits).

Perhaps unsurprisingly, skydiving safety statistics are not the easiest things to get hold of. Although fatalities are well-documented, to get a sense of the overall risk we need to know about all the times everything went to plan. After some searching, I decided to settle on a (regularly cited) statistic specific to parachuting in the US, where it's claimed there is approximately 1 death per 150,000 jumps.

150,000 is quite a big number, but how big is 'big'? The simplest way to look at this is to compare it with other travel safety statistics. After all, we don't tend to worry too much when getting on a bus or a train, or even into our cars, so how does skydiving compare? It sounds easy, but unfortunately when travel safety statistics are involved, things aren't entirely straightforward.

"Planes are statistically the safest way to travel" is a familiar mantra for those who get a touch nervy when heading to the airport. But what does this actually mean? It turns out there are (at least) three different ways one can define 'safest' - fewest deaths per journeys taken, time travelled, or distance travelled - but only one of these puts air travel at the top.

Travel safety statistics - deaths per billion...
 
...journeys
Bus: 4.3
Rail: 20
Van: 20
Car: 40
Foot: 40
Water: 90
Air: 117
Bicycle: 170

...hours
Bus: 11.1
Rail: 30
Air: 30.8
Water: 50
Van: 60
Car: 130
Foot: 220
Bicycle: 550

...kilometres
Air: 0.05
Bus: 0.4
Rail: 0.6
Van: 1.2
Water: 2.6
Car: 3.1
Bicycle: 44.6
Foot: 54.2 


Motorcycle: 1,640 Motorcycle: 4,840 Motorcycle: 108.9
Whilst the relatively large distances usually covered by aircraft mean they have a huge advantage over most other forms of transport (space travel is presumably an exception), in terms of deaths per journeys of time spent travelling it's buses and trains that lead the way. Which statistic one chooses does, as always, depend on what question one is interested in answering (and the relevant Wikipedia page has an interesting discussion of this), but it is perhaps worth noting that buses and trains are safer than cars on all three counts, whilst motorcycles are the most dangerous by a considerable margin.

So where does skydiving fit in? We already know that (in the US, at least) one sees approximately one death per 150,000 jumps. Converting this so we can compare it with our table, that's around 6,700 deaths per billion 'journeys'. Fairly dangerous then, but perhaps surprising that (in terms of journeys taken) a parachute jump is only about four times as dangerous as taking a motorcycle out for a spin.

Calculations for time and distance are trickier (and arguably meaningless), given the variation in jump height, angle of fall, chute deployment times, and so on. However, in the interests of completeness I'll assume an average fall of 14,000 feet (around 4.3 kilometres), a fall time of around six minutes, and (most importantly) that no-one is going to be too concerned with such a slapdash approach. These numbers work out to about 1,500 deaths per billion kilometres, and a whopping 67,000 deaths per billion hours travelled. Not very safe (although admittedly not very scientific, either). Guess my parachute commute will have to wait.

http://www.statslife.org.uk/science-technology/1360-sky-high-your-risks-when-taking-a-freefall
-------------------------------