Interesting article. You rail against Vox for misleading interpretations of data, but why are you doing exactly the same thing?
For example, finding a mathematical correlation isn’t what “More Guns = More Gun Deaths” is all about. For the developed world, America is held up as a warning example of what not to do. (Not just in gun deaths, but in a lot of other things as well, such as wise and coherent governance.)
Any First World nation can look at their statistics on gun ownership and gun violence and compare them to those of America.
In every single case, both are higher. Much, much higher. Like many times or even orders of magnitude higher.
Every single case.
And yet you say there’s no correlation at all.
That’s because you are giving every data point and every relationship the same weight. You’ve got Iceland with 355 000 population, and Luxembourg with 602 000 on that graph, and your model is pretending that they all count for the same.
It’s like saying that Peugeot and Maserati are just as important in the American car market as Ford and GM, and the relationship between price and accident rate of Maserati is just as valid as that of Ford.
Such small populations skew the rates. Some nut goes berko in Iceland, takes out his parents-in-law and then shoots himself, suddenly the gun death rate goes up by a factor of a hundred. An American-style school shooting and now Iceland is topping the charts for rates of violent death.
America is an outlier in gun ownership and gun deaths, and because there’s a third of a billion Americans, it’s a solid, significant, meaningful data point.
If you had run your article past a statistician, once they had stopped laughing, they would have pointed out the flaws in your argument.
Giving flawed results a veneer of statistical respectability is wrong. You do it, Vox does it. Both wrong.