It's still ridiculous to tie it to 'race' because, a) there is no real biological classification of race, which means all of this data using such classifications is already inherently biased, and b) mere correlation again does not entail a causation, so there could very well be a third variable that ties it all together. I can think of a bunch right off the top of my head, like stereotype threat and the still all too common racism resulting in many educational disadvantages for black students.