I'm surprised (maybe I shouldn't be) that no one on this post has mentioned omitted variable bias or the difference between bivariate and multivariate regressions as reasons correlation does not imply causation. Seems pretty basic to me.
UPDATE: Actually said before here, in the last paragraph, which I'll admit to skimming. I blame Google.
"Thou shalt not extinguish thine anger, but shall master it,
that thy conscience may not be blunted by adjustment
to wrong causes."
-The Dutch Ten Commandments to Foil the Nazis
No comments:
Post a Comment