,
8 tweets,
6 min read
Read on Twitter
@RAVerBruggen @alv9n Part of the problem here is a fundamental lack of appreciation from what we would expect to find even if students were perfectly randomly distributed, i.e., zero segregation as measured by entropy indices and the like.
@RAVerBruggen @alv9n In *most* states, the weighted black exposure to white students isn't more than maybe 10 percentage points from what we would expect to find if black and white students were randomly distributed in schools across the entire state.
@RAVerBruggen @alv9n Meanwhile, the effects of demographic change over the 30-40 years have been absolutely massive in many states.
@RAVerBruggen @alv9n Compare southern states in 1988 to 2016 (x-axis is % of state student population white, y-axis is weighted black exposure to whites). if random across state they should be on the black line (slope=1) 
@RAVerBruggen @alv9n In other words, It's not that there's been a strong secular trend for whites and blacks to pull further apart, but that demographics nationally have shifted to be sharply less white. Even if there had been no change in segregation, we'd expect a large increase in "isolation".
@RAVerBruggen @alv9n In a bunch of states, the student population is less than 50% white today (even lower in many broader metro areas). e.g. CA and TX are ~25% white, which implies on average weighted black exposure to whites should be around 25% if random distribution +
@RAVerBruggen @alv9n and probably somewhat lower realistically given that a good many groups live in different MSA/micropolitan areas, rural areas, etc.
@RAVerBruggen @alv9n Put differently, even in a world with zero entropy-style segregation, blacks in many states would still have *less* exposure to whites than actually existed in 1970 with measurable entropy segregation (or blacks in relatively segregated yet whiter states etc)
