Travel Safety Statistics

The relative safety of different modes of travel is something that most people have misconceptions about. They’re not helped by official statistics, indeed official statistics add to the confusion.

It’s not that the statistics are actually inaccurate, it’s that the impression they give is inaccurate. I don’t know whether the statistics are accurate or not, but I’m inclined to believe that they are. The impressions they give, and the beliefs many people hold as a result, are not actually supported by the statistics, even if the statistics are true.

The basic statistic involved is the “fatalities per passenger mile” figure – which is claimed to be lower for flying than for any other major mode of travel, probably truthfully.

On the face of it, “fatalities per passenger mile” looks like an appropriate measure of travel mode safety. If railways carry ten times as many passengers as airlines, and have five times as many fatalities, then other things being equal, they seem to be twice as safe. Other things aren’t equal though: passengers on planes on average travel much further than train passengers, so that’s a factor in the plane’s favour. But let’s examine this more closely.

What we really want to know is whether it’s safer to travel by plane or by train for the particular journeys we actually want to make.

We don’t have enough information here to make any kind of sensible judgement at all. Plane accidents are much more likely to occur near the beginning or (intended!) end of a flight, rather than while cruising at altitude between airports. Likewise, railway accidents are much more likely to occur at railway junctions or road intersections than on long stretches of uncomplicated track between cities. This means that the statistics of accidents per passenger mile, for both trains and planes, are strongly dependent on the length of the journeys. Shorter flights, with a larger proportion of their mileage close to the airports of departure and intended arrival, are less safe (per passenger mile) than long flights; likewise, short train journeys, with a larger proportion of their mileage in the cities of departure and intended arrival, are less safe (per passenger mile) than long train journeys.

If you’re travelling thousands of miles, in general you don’t have the option of travelling by train, and if you’re travelling ten or twenty miles, in general you don’t have the option of travelling by plane. It’s only for distances of a few hundred miles where you have both options – and those are the least safe flights per passenger mile, and the safest train journeys. Which is actually safer? You can’t tell from the simple “fatalities per passenger mile” figure. And that’s the only figure they’re going to give you.

You can’t drive from the living room to the kitchen, either; and you can’t (in general) walk twenty miles to work and back. But going a mile to school, which is safer, driving or walking? Simple statistics for accidents per passenger mile in cars versus accidents per pedestrian mile don’t tell you.

One has to give officialdom the benefit of the doubt, and suppose that this confusion is due to their incompetence, but it’s sometimes hard not to think the obfuscation is deliberate. Airlines are the main beneficiaries of these particular misconceptions.

You have to look at statistics carefully. The implications are often not what they at first appear to be. This may or may not be deliberate misrepresentation; sometimes whoever is presenting the statistics may themselves misunderstand the implications. Very often the information you really want is actually not possible to deduce from what is presented, as in this case.

Statistics, if properly done, can be very useful in assessing relative risks – but they’re not often properly done, and even less often properly presented.