Thanks for the comment. Some clarifications.
You are absolutely correct that a multivariate analysis is more appropriate. But there are three things to keep in mind about that.
1) Media sources (Vox, Mother Jones, etc) are presenting bivariate analyses that are wrong, and intentionally misleading, and they crop up everywhere. One hit the front page of Medium literally today. https://medium.com/s/story/everything-ive-learned-about-gun-control-aka-holy-shit-stop-trying-to-ban-assault-weapons-27131ca4f05f
2) if you need a dozen control variables to show a correlation, then the most important thing to take away from that is that there are a dozen related quantities, possibly causes, to consider. Not one.
3) I embraced the most well produced multivariate analysis on the problem in the article, and explained some suspect methodologies, as well as how the conclusions from it are not being used properly in the media. You do not see Mother Jones trotting out headlines saying “Guns are one sixth as big a problem with homicide as black people.” There is a deeply misguided drive within the media to willfully misinterpret what these studies actually conclude.
Also, for what it’s worth, the CDC is not at all prohibited from doing research on guns. They can do all the research they like. They are prohibited from advocacy. And that prohibition was instilled because they were brazenly cooking the data to try and push an agenda, and were being relatively public about it.
Why Congress stopped gun control activism at the CDC
History is not always written by the victors. In a Nov. 20 letter to the Labor, Health, Human Services, and Education…
There are, of course, people on both sides of that discussion, as there always are. This isn’t a bad review: