Right we have studies, and they are 30 page documents that only academics read, and so they get poorly summarized by articles that try to make them sound more noteworthy than they are. Usually by saying "X linked to Y" without establishing causality or even if that link is a significant risk (is it a .1% increase in risk?)

When it's a drug more than 10% of the US population uses, we can immediately say the risk increase can't really be that big or we'd have noticed it by now.

Edit: after looking at the paper, it looks like among the weed group the prevalence is roughly twice as high -- so instead of 1/100 having psychotic issue it'd be 2/100... and again for people who used when they were 13-17 year olds, which is underage in every state.

So you could frame that as doubling the risk OMG, or a 1 percentage point increase in risk, or it could all just be self-medicating, we really don't know much. Probably still safer than alcohol.