The funny thing about statistics is that they can be highly misleading, especially when sample sizes are too small and the statistics are detached from their wider context.
There is perhaps no better example of this than the Concorde. It used to be, statistically speaking, the safest passenger plane type in the world, with zero accidents. Then overnight, it became the least safe passenger type in the world (and by a rather wide margin at that). Because of one single accident.
The thing is, aviation safety statistics are usually determined by seeing how many accidents a particular plane type has suffered in relation to the total mileage flown by that airplane type. (The idea here is, of course, that if a particular airplane type has flown, in total, let's say, ten million miles, and has suffered ten accidents, it would be unfair to consider it as unsafe as another airplane type that has also suffered ten accidents, but only flown one hundred thousand miles in total.)
While other passenger airplane types have typically hundreds or even thousands of units in service at any given time, and they fly constantly, their grand total mileage being staggeringly high, the Concorde had a ridiculously low amount of units (20 were ever built during the entire existence of the aircraft type) and flew a ridiculously small amount, compared to your typical passenger airplane. For this reason one single accident immediately skyrocketed it to the very top of the list of least safe passenger aircraft types.
If you only looked at the list after the accident, you would easily believe that the Concorde is extremely dangerous. But this isn't true. The statistic is skewed due to a too low sample size. We have to look at the details, the context, to see whether we can make that assessment or not. This particular statistic is not very good in this case because of the problem of sample size. (In statistical terms, the error margin is way too big to draw any conclusions.)
When American conservatives want to defend their gun culture, they just love to play the statistics game.
How to prove that the United States is not any more dangerous than other countries? Perhaps the number of gun-related crimes? No. Murder statistics involving firearms? No. Number of mass shootings? No. Number of mass shootings per capita? Still no. Oh, I know: Rate of deaths by mass shootings per million people since the beginning of the year 2000.
That's exactly what some American conservative websites are doing (and pro-gun conservatives parroting).
When we look at the number of deaths, per capita, of each country by mass shooting incidents (only mass shootings, not all gun-related deaths), the United States goes way down the list. A perhaps rather surprising country comes up on the top of the list: Norway.
Does this mean that there's an unusually high number of mass shootings in Norway, and it's an extremely dangerous country, and that you will be more likely to die from mass shooting in Norway than in any other country?
No. Like with the Concorde, that statistic comes up because of one mass shooting incident. The recent infamous one where something like 93 people died. The problem, with respect to this statistic, is that Norway has a relatively low population, about 5.2 million people. The United States, on the other hand, has almost 63 times that much population. Thus, when you calculate the number of deaths by mass shooting in proportion to the population, the result goes way down for the United States because of its massive population.
But, like with the Concorde, the sample size is way too low to make any conclusions. The statistic consists almost solely of that one single incident, and thus isn't very representative. You are still much more likely to be shot in the United States than in Norway.
The funny thing about statistics is that by choosing your methodology appropriately, you can make them say whatever you want.
There is perhaps no better example of this than the Concorde. It used to be, statistically speaking, the safest passenger plane type in the world, with zero accidents. Then overnight, it became the least safe passenger type in the world (and by a rather wide margin at that). Because of one single accident.
The thing is, aviation safety statistics are usually determined by seeing how many accidents a particular plane type has suffered in relation to the total mileage flown by that airplane type. (The idea here is, of course, that if a particular airplane type has flown, in total, let's say, ten million miles, and has suffered ten accidents, it would be unfair to consider it as unsafe as another airplane type that has also suffered ten accidents, but only flown one hundred thousand miles in total.)
While other passenger airplane types have typically hundreds or even thousands of units in service at any given time, and they fly constantly, their grand total mileage being staggeringly high, the Concorde had a ridiculously low amount of units (20 were ever built during the entire existence of the aircraft type) and flew a ridiculously small amount, compared to your typical passenger airplane. For this reason one single accident immediately skyrocketed it to the very top of the list of least safe passenger aircraft types.
If you only looked at the list after the accident, you would easily believe that the Concorde is extremely dangerous. But this isn't true. The statistic is skewed due to a too low sample size. We have to look at the details, the context, to see whether we can make that assessment or not. This particular statistic is not very good in this case because of the problem of sample size. (In statistical terms, the error margin is way too big to draw any conclusions.)
When American conservatives want to defend their gun culture, they just love to play the statistics game.
How to prove that the United States is not any more dangerous than other countries? Perhaps the number of gun-related crimes? No. Murder statistics involving firearms? No. Number of mass shootings? No. Number of mass shootings per capita? Still no. Oh, I know: Rate of deaths by mass shootings per million people since the beginning of the year 2000.
That's exactly what some American conservative websites are doing (and pro-gun conservatives parroting).
When we look at the number of deaths, per capita, of each country by mass shooting incidents (only mass shootings, not all gun-related deaths), the United States goes way down the list. A perhaps rather surprising country comes up on the top of the list: Norway.
Does this mean that there's an unusually high number of mass shootings in Norway, and it's an extremely dangerous country, and that you will be more likely to die from mass shooting in Norway than in any other country?
No. Like with the Concorde, that statistic comes up because of one mass shooting incident. The recent infamous one where something like 93 people died. The problem, with respect to this statistic, is that Norway has a relatively low population, about 5.2 million people. The United States, on the other hand, has almost 63 times that much population. Thus, when you calculate the number of deaths by mass shooting in proportion to the population, the result goes way down for the United States because of its massive population.
But, like with the Concorde, the sample size is way too low to make any conclusions. The statistic consists almost solely of that one single incident, and thus isn't very representative. You are still much more likely to be shot in the United States than in Norway.
The funny thing about statistics is that by choosing your methodology appropriately, you can make them say whatever you want.
Comments
Post a Comment