Fact: Eating margarine in Maine does not cause you to get divorced.
Maybe this week’s fact seems like an obvious one. I’m guessing that the question of whether eating margarine causes couples to divorce anywhere in the world was one that was far from the tip of your tongue.
Why am I bringing this up? Well, just look at this compelling graph:
Its undeniable – you’d be a fool to argue that the two aren’t related, but common sense dictates that of course they aren’t (unless there is something very strange going on in Maine). That strange paradox is the critical difference between correlation and causation.
Causation is when one event causes another event to occur. For example, drinking beer makes you feel drunk. Correlation is when two things both happen at the same time, but there may be no link between the two: it may just be a coincidence that they happened at the same time. For example, if you sneeze and there is suddenly a power cut: it could be that your sneeze has caused that power cut, but it’s more likely that the two simply happened together by chance.
Sometimes we see a correlation because one thing does in fact cause the other, like when you drink a beer (or several) and get drunk, and in the examples above it’s obvious which events are correlations and which show causation. It gets more difficult when we’re not sure whether one thing causes another, and in fact, determining whether the results of experiments are just a correlation or a there is meaningful causation is one of the greatest challenges scientists face.
Check out Spurious Correlations for more hilarious correlations that absolutely do not imply causation.
‘Another One of Mr Shaunak’s Little Bites of Science’
Image under Creative Commons License