May 21, 2014

Post hoc ergo propter hoc

Just because there's a correlation between two data sets doesn't automatically mean there is causation between the two. Take, for example, the various examples of spurious correlations at the blog Spurious Correlations. The site will even generate new correlations at your request.

correlation = -0.933

correlation = 0.992

 correlation = 0.993

These aren't from Spurious Connections, but they're relevant, too.

From Dinosaur Comics
From Glen Gilchrist
PS: In retrospect, maybe the title should be cum hoc ergo propter hoc rather that post hoc ergo propter hoc.

No comments: