[haihaibye]

If groups perform differently on tests vs grades, different groups will lose or gain depending on the relative weightings between the two.

The best thing about having 2 measures that don't agree is you can pick which one you want to be the true source, and which one is terrible and wrong. Then you can write a story that fits your desired narrative.

If disadvantaged minorities did better on tests vs grades, do you have any doubt we wouldn't have a scandal about teacher's highly subjective racist marking?