We’ve done all sorts of dumb things for thousands of years.
It’s one metric of many. We know that paying for a tutor can change test scores. We know that a shitty home life can, too. They’re just harder to measure.
If you’re suggesting that having “a shitty home life” can make people perform badly on tests, but not perform badly in real world tasks, we don’t “know” that. It’s something people want to be true, but there’s not much evidence for it. Meanwhile, there’s reams of evidence that standardized tests scores are highly predictive of performance.
Okay, but where is the evidence that those people perform better in real world tasks than their test scores would indicate once they are no longer in that shitty home life?