Hypothetically, what if it was true that wealthy people's exposure to better education over a lifetime, reduced stress distracting them from learning, better and more abundant roll models, etc, meant they would do better on pretty much literally any test you could put in front of them?
Then wealthy people show more "merit" and the test is working as designed. If the test is an accurate predictor of success in college and success thereafter (which is a big if), then that people who do well on it are more deserving of college admissions. The ancillary qualities of that group of people (wealth, race, etc.) are irrelevant.
The SAT is (well, was) supposed to predict academic success at a college/university. The idea being that an institution could admit students who would likely graduate, as opposed to students who did not.
This is the reality of the situation that people don't want to admit. This is a social engineering problem related to the outcomes of college entrance processes. It's not really about the test or how we are measuring aptitude.
This problem would be far better solved elsewhere in the chain instead of blurring the meaning of data and pretending that things are different.
Yes but fucking with the metrics costs almost nothing. Does almost nothing too.
It can even backfire by sending a bunch of disadvantaged kids to programs they're not prepared to succeed at to "help them" get saddled with college debt they can't discharge.
Most research points to this being true.