Burying the lede

Well this is full of surprises. A press release regarding research on arrests from the University of South Carolina, Half of black males, 40 percent of white males arrested by age 23 by Peggy Binette.

Key findings from the report.

• By age 18, 30 percent of black males, 26 percent of Hispanic males and 22 percent of white males have been arrested.
• By age 23, 49 percent of black males, 44 percent of Hispanic males and 38 percent of white males have been arrested.
• While the prevalence of arrest increased for females from age 18 to 23, the variation between races was slight. At age 18, arrest rates were 12 percent for white females and 11.8 percent and 11.9 percent for Hispanic and black females, respectively. By age 23, arrest rates were 20 percent for white females and 18 percent and 16 percent for Hispanic and black females, respectively.

The press release makes much ado about the differences between arrest rates by race but I think that is substantially a red herring, the imposition of pre-existing blinkers on a new dataset which allows you to arrive at conclusions already formed.

Here are the surprises to me.

* Magnitude - 40% of American males have been arrested by age 23? Wow! Could that possibly be right? Is there a flaw in the methodology somewhere? Perhaps. But if these data are correct, I suspect there is something much more insidious behind the averages. I suspect the numbers are far more variant by class (using income quintiles as a proxy) than by race. I suspect that the percentage of boys in the bottom two income quintiles would be staggering and those for the top two quintiles minor.

* Race disparity - Not nearly as significant as I would have anticipated. Not to say that the gap isn't material, 38% versus 49% for white males versus black males, just not as large as I would have expected (and I would have expected both of them to be much lower).

* Gender disparity - A little larger than I would have expected, basically 12% for female 18 year olds and 25% for males. This is before college and before adulthood. Yes, it is anticipated that males are greater risk takers, more physical, and more likely to break rules. But twice as much at such a young age?

* Race inversion- Not discussed at all in the press release is what I find most surprising of all. While female arrest numbers are very close by race at both 18 and 23, in both instances it is white women who have the higher arrest rates. What's going on there and how does that fit into the received racial narratives? I don't know, but I think it is notable that the most counterintuitive finding is the one that is not discussed at all. I am guessing that the measured differences are within the margin of error, i.e. that all females basically have the same arrest rate given the margin of error. Still, it would be nice for them to address that.