"Why do people love to say that correlation does not imply causation?"
Daniel Engber at Slate observes that use of the phrase seems to have skyrocketed in the past 50 years and he speculates that it's because we are now more concerned about false positives than false negatives.
I need to think about his conclusion, but he makes sense on misuse of the phrase. Which means I've been misusing it. Damn.
Also interesting: the reference to McCloskey and Ziliak and statistical significance. Something I'd never considered.
Posted by: Michael Greenspan | October 16, 2012 at 07:32 AM
Perhaps we need a counter-catchphrase, like "Correlation is correlated with causation."
Bruce Charlton supports Engber's idea that we fear false positives more than in the past, but offers a different reason why, in this post: http://charltonteaching.blogspot.com/2012/10/whats-going-on-in-world-look-out-window.html. He thinks the rise of media and the Internet has led modern intellectuals to conceive of themselves as straining vast amounts of information to find a few nuggets of truth. We have more low-quality information around than in the past, so avoiding false positives is more important.
Posted by: Paul Jaminet | October 17, 2012 at 07:43 PM