3 ms·

I would say specifically that controlling for other variables is tricky when there are a lot of them that strongly correlate: it's very easy to still have a res

I would say specifically that controlling for other variables is tricky when there are a lot of them that strongly correlate: it's very easy to still have a residual effect after trying to remove them. So if those other factors are much stronger effects, it should increase the standards of evidence for smaller factors.

2muchcoffeeman 2y ago
“Correlation does not imply causation” has become a catch phrase that people whip out just to disagree. It’s fine to disagree, but at least write a sentence to say why. If you can’t say why, then just take it as an interesting anecdote.