Monday, 29 June 2015

Does cheese consumption cause death by bedsheet? Correlation is not causation

Yes, there is an apparent perfect correlation between Per Capita Consumption of Cheese vs Number of people who died by becoming tangled in their bedsheets.

This and other crazy correlations come from the research I am doing on 'Evidence-based Decision Making' for the Knowledge and Innovation Network Autumn Workshop. I was reading an article in this month's Wired magazine about 'spurious correlations' when I came across Tyler Viglen's website. Tyler has lots of funny and ridiculous examples to prove that correlation is not the same as causation. Even when decisions or inferences are made on the back of 'evidence', it's worth understanding what you are looking at.

Here is Viglen's random correlation generator http://tylervigen.com/discover. Have fun.
Related articles

No comments: