Given the discussion is racism here, wouldn’t it be useful to note that over 91 percent of blacks are murdered by other blacks. Similarly, 84 percent of whites are murdered by other whites. One thing is clear, whites aren’t getting guns to kill blacks…..
…..Now rerunning the same regression on non-whites. Despite the much smaller sample, you still get statistically significant effect for this supposed measure of symbolic racism. So are non-whites who are racist against blacks more likely to own a gun? Hardly seems reasonable.
If you are seriously into statistics, you will enjoy his careful dissection. I will admit that this is a wee bit over my head, but Lott’s has a long tradition of bulletproof research so I’ll go with him.
The Cartoon Guide to Statistics is actually an excellent way to become functionally literate in statistics without having to get seriously into them.
Added to the list.. Thanks!