After blogging and writing a diary yesterday regarding some work I've done in measuring diversity in populations in different nerdy ways, I thought I would be done with the subject for a while. However this morning's news of the Michael Brelo acquittal got me thinking again.
I thought at the very least I should take a look at the Cleveland Police Department. Which I did. The short unnerdiness of it: There are massive deviations between what the population of Cleveland looks like and what the Police Department looks like. The simplest calculations of this is that African Americans make up 53% of the general Cleveland population, but only 27% of the police force.
I calculated my "diversity index" on this, and found a relatively (when you look at other institutions like the US senate) mild deviation of about 1.9 (police) to 2.5 (population). This is largely because the diversity index doesn't perform well under a condition that I refer to as "the apartheid problem."
For all the nerds out there, I also played with using what's called a "goodness of fit" test to measure the differences between Police Departments and underlying populations. With some success.
Here's a link to my full analysis.