To smile...

To smile (and more…) : correlation is not causation

13 May 2014

Tyler Vigen, a young scientist who works for the US army, has published a huge number (>17000 !!!!) of « spurious correlations ». In fact, it’s real correlations, but without any signification.

Per capita consumption of cheese (US) correlates (Correlation: 0.947091) with number of people who died by becoming tangled in their bedsheets

Age of Miss America correlates (Correlation: 0.870127) with Murders by steam, hot vapours and hot objects

Tyler Vigen continues to publish new correlations each day.

1405SpuriousCorrelations

In statistics, correlation is a statistical relationship between two variables or two sets of data.

It could be a indication of a causal relationship. But it also could be NOT! Voir “correlation” article on Wikipedia

A purported causal relationship between two correlated sets of data is often used by scaremongerers.
At the oppposite, in science, the interpretation of a correlation, even  if it’s a strong one, must be very cautious