[viraptor]

> These studies are correlative

Once you have population studies from millions and include separated/not twin studies to exclude environment differences, you really need strong arguments and theory why the inheritance is not the mechanism here.

> but that doesn't mean the counterfactual is also true: "everybody with these genes will have ADHD, no matter what".

Nobody claimed that. That's not how genes work. That's not how any(?) inheritable disease works. The whole paper is about quantifying the chances. Even the most basic thing like eye colour has a random element to it.

> if that were true then the DSM would describe a genetic test for ADHD

There's no simple mapping and the candidates for complex interactions are in 200+ range the last time I've seen. There's no "you have this one gene, you have ADHD" and nobody claimed that. The work on narrowing down how exactly the candidates interact / get expressed is still ongoing, so hopefully one day we will have a genetic test.

[tdpvb]

Right, so then "ADHD genes" are only correlated!

(You'd said, "it's [ADHD] very highly inherited" -- which sounds declaratively causal. But if you didn't mean this as a causal statement, then cool. No worries.)

[viraptor]

It depends if you want to be ackshualy technically correct or conversation correct. We won't confirm causality until we flip people's genes to give them ADHD or describe the whole exact interaction at whole-organism level, so practically... never.

On the other hand, for normal purposes, we know it's inheritable. You'd need to discover some whole new mechanism for how diseases could possibly transmit to prove otherwise. There's enough real world data to say this.

[tdpvb]

Conversational or otherwise, you keep saying it's inheritable but these "ADHD genes" are only correlated with increased risk, not predictability. That's key, because one is nature, the other nurture.

Also, proving a correlation doesn't necessarily mean it's exclusively, automatically the only variable. That's the tricky bit with confounding variables: they could be one of several, or many, correlative effects, and where none could be the cause -- much less a useful avenue of solution. It's helpful, but not conclusive.

Great, we found some genes. But other studies should also look at psychotherapy, or even diaphragmatic breathing (asthma is another great example). Alternative analysis, i.e. anything other than the most lucrative option for pharmas to sell drugs -- as in, alternative solutions that are closer to the root cause, not just the most complicated.