[tlb]

Typically, knowing the mechanism comes long after solid evidence of the effect.

The mechanisms by which cigarettes cause lung cancer are surprisingly complex and weren't fully understood for a long time, but it's obvious that some kind of mechanism is plausible.

So usually the bar for taking something seriously is: statistical evidence of a significant harmful effect, and at least one plausible hypothesis for a mechanism. But you don't have to be sure of the mechanism.