Jonathan Haidt is someone who nobody should take seriously. Pretty much all of the data he cites is cherry-picked and the vast majority of people in trust and safety and similar will tell you that he is probably one of the least reliable authorities on this subject. He's aiming to sell fear, not to actually solve the problem.

The only data point I referred to is the dramatic rise starting in 2015, and the divergence by gender. I don’t think anyone seriously disputes that, just the presumed usage patterns he thinks go along with those. The criticisms of him are basically orthogonal to the point I’m making.