Certainly no Nazi. However, his issue is that he is apparently unable to distinguish between main effects of populations and diversity effects (this is an interaction). They are saying the outcome is explained by the main effects, not the interactions.
-
This Tweet is unavailable.
-
This Tweet is unavailable.
-
The key to understanding it is that a given area can have high diversity for multiple reasons. An area with a 50-50 White-Asian split has high diversity, but is not expected to perform poorly because composite groups perform well.
Compare with an area with 50-50-Black-Hispanic mix. This will perform poorly because the constituent groups do so, but it has the exact same amount of diversity as the 50-50 White-Asian one. Statistically, thus, one can work out which predicts what.