Although the title is not a phrase that the average man in the street will come across too often, it’s a favourite amongst scientists and statisticians. It’s a term that these experts use to stress that just because there is a connection between two things does not automatically mean that one caused the other.
A simple dictionary definition of correlation is that it is a ‘mutual relation of two or more things’. As there is a need for there to be a relationship for there to be a correlation, it means that there can be no causation (a reason why something happens) without correlation.
Where the idea that correlation prompts or proves causation falls down is that it assumes that there is only one correct interpretation of the facts.
This is known as a logical fallacy, a situation where two events happening at the same time is said to prove a cause-and-effect relationship.
In simple terms, the idea of ‘correlation proves causation’ suggests that if event X occurs in correlation with event Y, then event X must have caused Y to happen.
With just a bit of thought however, it should be plain that there could be many different ways of interpreting this information.
Of course, it may be that X causes Y but it is equally possible that it is the other way around, that it is Y that causes X.
It’s like a more sophisticated version of the old chicken and egg debate – which came first?
Then there might be a scenario where both events X and Y were caused by an outside party. It is the outside element that is responsible for what happens, nothing to do with either X or Y.
For instance, it is true that as ice cream sales increase, the number of children who drown also rises sharply.
Following the ‘correlation proves causation’ argument, the logical conclusion would be that eating ice cream causes drowning. This is of course nonsense and completely ignores the truth that ice cream sales rise when the weather is hot which is when children are most likely to go swimming.
Then there are scenarios where the correlation between the two events is so distant that it is little more than coincidence. If it happened to be your 5th birthday on the day Neil Armstrong first landed on the moon, it doesn’t mean that his landing occurred because it was your birthday, does it?
So, there you have it. Correlation is not equal to causation and don’t let anyone tell you otherwise!