Which seems like it just is saying that the proportion assigned to gifted programs was lower than what the average test score would suggest (notable when compared to hispanics), but hispanics could have a wider distribution (or a bimodal one), so this seems like a shoddy test. The correct check of the hypothesis would be to find the test score threshhold above which 50% of students get into the gifted program and see if this threshhold varies by race... but I can't find anywhere this is done.
Trying to chase it down just got me to this: https://cdn.vanderbilt.edu/vu-news/files/20190417222934/Gris...
Which seems like it just is saying that the proportion assigned to gifted programs was lower than what the average test score would suggest (notable when compared to hispanics), but hispanics could have a wider distribution (or a bimodal one), so this seems like a shoddy test. The correct check of the hypothesis would be to find the test score threshhold above which 50% of students get into the gifted program and see if this threshhold varies by race... but I can't find anywhere this is done.