A website set up by a student at Harvard teaches us to look carefully at statistics. And it’s fun at the same time.
“Margarine consumption linked to divorce.” If you saw that headline on a newspaper or website, what would you think? What if you read a little further and found a compelling graph showing the rates of divorce and margarine consumption tracking each other closely over almost 10 years. Tempted to believe there could be a link?
“Maybe when there’s more margarine in the house it’s more likely to cause divorce,” muses Tyler Vigen, “or there’s a link with some of the molecules in margarine or something.”
Vigen is the man behind the margarine graph, which he published on his website Spurious Correlations. The name gives the game away – he’s a statistical provocateur.
“I’ve seen a lot of headlines, especially sensationalist ones – ‘Scientists find a connection between x and y,’ he says.”
“In a lot of those situations there might be a correlation, but it’s really important for us to be critical about whether there’s a causal mechanism.”
One of the golden rules of statistics is that correlation does not equal causation. Just because the movements of two variables track each other closely over time doesn’t mean that one causes the other.
To make this important, but somewhat dry, point more accessible, Vigen, a criminology student at Harvard Law School, wrote a computer programme to mine datasets for statistical correlations. He posts the funniest ones to Spurious Correlations. …
The site contains plenty of raw material to test out your ability to come up with a creative causal mechanism.