There might also be some risk of psychological damage when you use a criminalized substance that has anxiety as a common side effect. For example, it's easy to see how that could make you a bit reclusive.

So anyone who's really concerned about the mental health of cannabis users is obliged to consider the damaging effects of its very criminality.

(There are all kinds of other factors. Like cannabis use being correlated with higher trait openness, which has its own particular correlations with certain personality disorders, et cetera.)

Not to mention that such criminalized substance might well have been mishandled by criminal resellers with no regard for public health. Any time one buys a criminalised substance, it's like he was going to eat a box of unlabelled chocolate: maybe it's safe to eat, or maybe it's full of rat excrement - you don't know until it's too late.

It might "just" be a correlation but it might be heavily interconnected. If you have data proving your "just correlation", please provide them.

Correlation without causation is a lack of data — the burden of proof is on the claim not the questioning of it

What you have mentioned amounts to correlation only. If it's also causation, you're the one that should provide the proof.

I suspect this is a big factor. Self-medication is a real thing. When people don't feel good, they start looking for things to make them feel better. And adolescence is a time when a lot of people don't feel good.