> I think is clear that the baseline of any study is that correlation does not imply causation. You start there. That’s a basic intelligent start. You never take for granted that one thing causes the other. The ideas is to try to identify if there are any potential paths for causation.

You might start there, but an incredible number of people read the headline and think "oh man; I guess I gotta increase my iron intake".

It's said a lot because it's so easy to draw a false conclusion from correlations.