On the conflation of correlation with causation

TPM Cafe contributor FlyOnTheWall has an excellent post (as per usual) on "Signals, Noise and Polling." It is very insightful and totally correct, whatever one thinks of Bittergate. I won't excerpt it; just, by all means, read it.

Comments

[Condor]

I always had a feeling that correlation and causation went hand in hand. Now I know that it's because causation affects correlation.

[Andrew]

I mostly don't disagree; the TPM analysis isn't that far off from the National Review blogger's take.