[demallien]

Noooooo, that's the whole point! The question that is being asked is not "how to make gun owners less likely to shoot people" but rather "how to make it less likely that people get shot". Per capita is absolutely the right measure to use for the latter question.

[kromem]

In terms of rampage shootings, the way to reduce the number of shootings is to reduce the number of guns. That's what the data shows.

A per capita analysis starts discussions about how the US is different in multiple ways. Oh, well, in the US a person is 5x more likely to go on a shooting rampage than in other OCED countries. Maybe it's US media coverage? Maybe we don't have the background checks or psychological evaluations other countries have. Maybe it's our action movies?

No.

It's simply that we have more guns. In fact, on a per gun basis, other OCED countries combined have a 50% higher rate of spree shootings. We just have a LOT more guns than other countries.

Now, whether this is a good or bad thing is a much broader discussion. But the ONLY data driven answer to "why does the US have more spree shootings than the rest of the world?" is: "Because the US has more guns".

TL;DR: A per-capita analysis leads to sociological problem solving as the base metric is people-based. A per-gun analysis cuts directly to the significant factor.

[maxerickson]

I wonder how good a proxy the number of guns is for the number of people with access to guns.

(It's probably reasonably good, I'm interested in what the numbers would be, not trying to put forward a conclusion)

I guess in the US over recent decades the number of people with access to guns has increased relative to the number of guns (because I think there are less hunters and more self defenders).