Hypothetical: if teacher A has a class of poor black students, and teacher B has a class of rich asian students, shall we directly compare test scores? If teacher A's students perform 1 stddev worse than teacher B's, shall we fire teacher A for poor performance?
If you don't want to do it for teachers, why do it for educational systems?
I'm sure that's how you see it, but the fact is, that the OP posited the idea that unions were incompatible with a healthy education system. There are simply too many counter-examples to that.
Further, you've pushed the conversation into a discussion about race and class by way of an abstract hypothetical. From your posting history, I'm not really interested in having such a complex discussion with you.
If you don't want to do it for teachers, why do it for educational systems?