Ok. I just really read all the stats here. I was only skimming them before, because as I said repeatedly, they aren’t really a part of my argument…. and frankly I trusted Erin, Dmitri and Michael to get them right. Which they did. Also hi Dana.
Now that I’ve had a chance to go through them I can say a real problem here. And Vic, you’re really not going to like it…. ignoring the racial aspect or the specifics of the conversation…. you’re getting a failing grade in basic statistical analysis here. I mean it… and I’m not an expert. I usually outsource mine to Stephanie Siler when I need them done. Because she is an expert… but I do know the base concepts and you’re just simply doing them wrong. And I expect you like your “results” because you think they confirm your hypothesis. They don’t. Strauss told you several times that you’re numbers support HIS claim…. not yours. He’s right.
Here are the points you’re missing:
1) you aren’t normalizing population sizes. (There are more non black people in america than black people)
2) you aren’t comparing the right numbers.
3) you don’t seem to understand the difference between correlation and causation. Namely, statistics can never really prove the latter. For anything. They can support hypotheses about causation when complex bi directional studies are set up. You don’t know how to do that. Hell, I don’t know how to do that. Which again, is why I trust steph. But even then, they just support, not proves…. because statistics are about correlation and not causation
4) I don’t think you realize that statistical significance is a thing…like it’s not just enough to say 99.99% of black men won’t be killed by police and 99.999% of white men won’t and that’s “basically the same” even if your math was right. Once you actually determine the actual percentages, you have to do a regressions analysis to determine significance. It can’t just “feel right”. Significance means something.
5) because you screwed up #2, you can’t just do number #4. What you actually want is to prepare the deaths per crimes per arrests per population of black people vs the same stat for white people and then p-test the significance between those two numbers. And that will tell you the significance. It’s not just enough to say it “feels the same” the way you are…. it’s like an actual mathematical concept that means something.