Correlation and Causation
22 February, 2013 1 Comment
(This is an adaptation of a Swedish blog post I wrote a couple of days ago.)
People who are in favor of gun control probably don’t read this blog. It is simply too obvious that criminals prefer disarmed victims to armed ones. It is equally obvious that dictators and would-be dictators prefer a disarmed population to an armed one. Not understanding this is stupid – and I wouldn’t insult my readers by assuming they are stupid, would I?
But there is a certain argument one hears once in a while: If one points to statistics that confirm that fewer crimes are committed when people are allowed to wear arms, one will hear that this is an invalid argument, because “correlation isn’t causation”.
Now, correlation isn’t necessarily causation. Some statistical correlations are mere coincidence. At best, correlation shows that there some causal connection to look for. To take a familiar example, smokers contract lung cancer more often than non-smokers. That means one should look at the possibility of a causal connection. On the other hand, not every smoker contracts lung cancer – and sometimes even non-smokers contract lung cancer. What this implies is that there is a causal connection to look for, but that there are also other factors involved.
On the other hand, causation always and necessarily implies correlation. For example, the law of gravity implies that the vast majority of things one drops will fall to the ground, and if some things do not fall to the ground, one will have to look for the factor that counteracts gravity. If a leaf or a feather does not hit the ground, or does it with a delay, this is explained by air resistance. If birds and airplanes manage to fly, it is because of a cause that counteracts gravity, and therefore birds fall to the ground only when they are shot, and an airplane (or a space ship) only when its motor for some reason stops working properly.
And in the case of gun control failing to avert crime and even causing crime, the cause is equally obvious. Criminals may be intelligent or stupid, but they are never so stupid as to choose armed victims when there are disarmed victims to attack. Neither would a dictator, or a would-be dictator, fail to realize that an armed citizenry would be a formidable obstacle. Not even a madman on a shooting spree would miss this.
It has been pointed out, time and time again, that mass shootings in the vast majority of cases have been perpetrated in “gun free zones”, and that they never occur at shooting ranges where every potential victim is already armed. It has also been pointed out that crime rates rise when guns for self-defense are outlawed. It is simply ridiculous to say that this is a mere statistical coincidence and that it is a correlation that does not prove causation. We know the causation and should not be surprised by the correlation.
Even proper ideas can be misused, and this is an example.