I see them mention a change in that study and that they did some binomial regression for ethnicity in those years. But it seems that's for secondary attributes like ethnicity, not for the main cause of death (suicide). Or am I missing something?
I'm aware of those examples. Those examples were before the internet, meaning people commit suicide they way they know how. Socially, you couldn't just ask someone how to commit suicide with an alternative method, but now you can look it up on online so you don't have to ask anyone.
The coal gas example does show a reasonable reduction (18-13). But after the conversion process was completed, you see it's still creeps back up to 15 just within a couple years.
The paracetamol example might actually go against your argument. We see that the suicide rate overall is very steady after the package size change, which suggests that people found other methods.
I'm aware of those examples. Those examples were before the internet, meaning people commit suicide they way they know how. Socially, you couldn't just ask someone how to commit suicide with an alternative method, but now you can look it up on online so you don't have to ask anyone.
The coal gas example does show a reasonable reduction (18-13). But after the conversion process was completed, you see it's still creeps back up to 15 just within a couple years.
The paracetamol example might actually go against your argument. We see that the suicide rate overall is very steady after the package size change, which suggests that people found other methods.