The same way the entire US educational establishment officially groups and tracks students for the purpose of outcome comparison: by self-identified racial classification primarily, and by country of origin for immigrants secondarily. Was it not clear from my original comment, where I talked about white Americans and immigrant students?
It’s not so much unclear as premised on a number of false distinctions that make it unfalsifiable. For example, at what point and under what conditions does a foreign-born child of Polish immigrants that immigrate to the U.S. become white such that they move from one group to the other? How would you correlate however said group membership requirements correspond to test performance? Why does your curiosity about the relevance of Simpson’s paradox disappear when discussing the category of immigrants, a group with more diverse membership than any of your racial categories, and thus more probable to being prone to its effects?
That's what I was thinking given it's America, it just makes no sense in relation to the rest of your comment. If you want to know how COVID effected students, the most natural grouping is by school policies with regard to COVID.
