It's weird how there's a whole thriving genre of studies that look at things with incredibly complicated causal graphs, maybe control for a random smattering of

It's weird how there's a whole thriving genre of studies that look at things with incredibly complicated causal graphs, maybe control for a random smattering of confounding variables, find a correlation between some variables X and Y, and then declare victory -- even though what they found is seldom an actual causal relationship between X and Y, both in theory and in practice [1]. The researchers here at least had the honesty to be forthright about this, but given the limitations of this kind of study I'm not sure I see the point in even doing them. [1] https://www.gwern.net/Causality https://www.gwern.net/Causality