So, it's possible that drinking increases your chance to make an impulse purchase, and that purchase might be a gun. But that's not causation.

It's not causation, and I'm not trying to imply it is. But it is a correlation, and that is interesting enough on its own.

If you read the article it basically states the same thing. It basically states that drinking can increase violence, and drinking can also lead to unsafe handling of firearms. So, the cause of violence here is alcohol.

I don't care about the cause of violence, I care about the correlation.

And the final nail in the coffin is the following quote: "The authors say their findings can inform strategies to discourage drinking and thereby decrease the likelihood of handgun-carrying among youth and young adults in rural areas." So, the goal isn't to decrease violence. It's to decrease handgun carrying among young adults.

Carrying handguns among young adults leads to increased violence. Having "decreased handgun carrying" as a goal seems completely fine to me as long as people still have a choice.

Also, a sample size of 2000 12-26 year olds? That's about 142 per age, which is not too bad, but in my opinion 12-17, 18-20, 21-26 are vastly different groups. They themselves state that 19-26 is where the most drinking/gun owning occurs, which makes sense.

That sample size seems completely fine. Which measure exactly isn't to your liking? Can you be specific about what n or P you'd expect?

You could say the same thing about owning a car. But if I told you that drinking is linked to purchasing a car, you'd think that's stupid.

Yeah, I think you misunderstand the study. Of course I'd be very interested in knowing whether people who drink are more likely to buy a car. Why would I think that's stupid? Explain it clearly to me.

But I think I understand - you don't want people to lose access to guns, so it's easier to declare the study to be "stupid" than to accept the correlation as reality. The problem is that the study isn't making any judgements or publishing any guidance - it just presents a correlation. You shouldn't reject studies because you don't like possible future implications.