re: “Relevant Statistics” Thread

quote:

---

I don’t follow your disagreement with the standard normalization.

---

It's not a standard normalization. Let me try again to explain why.

If all violent crime were included, *then* we could normalize by population. But since only the interracial portion of crime is produced, it is already normalized in a way. That is what my population of 100 example was intended to explain. I guess people didn't follow.

The problem is this: to determine whether something is "more likely" than something else, we have to ask more likely than what? Well, the obvious baseline is random chance. And given that white people make up much more of the population than blacks, random chance says more white people will be victims of random violence. Approximately 5X more likely, actually. So if something happens to white people 9X more than something else, we would divide that by 5, not multiply. Or, if you prefer, we multiply by the inverse. In other words, it happens 9/5 = 2.2 times more often than it would randomly. Not 45 times more often.