[cge]

To add to the other answers: at top universities, you can also have very unusual applicants. At Caltech we'd sometimes have (standard age) freshman come in with research publications, for example.

If you have the resources to review applicants individually, sometimes there are people who clearly stand out for reasons unrelated to test scores. I expect that in most cases, if they've taken tests, they have good scores, and it can make sense to generally have a threshold, but if you simply rank by top scores they are likely not the absolute best, as they would have been spending more time and attention on things other than studying for them.

[pxx]

It's actually super common to have research publications on undergrad applications now, so much so that I've heard that having such a publication is "table stakes" to be competitive at top schools.

Simultaneously, I've also heard that it's easier to get your name on a paper as a high schooler than as an undergrad working in the same lab. Some of this is because professors are incentivized to put random high schoolers on their papers -- when applying for grants, they're often asked what K-12 outreach they've done (https://new.nsf.gov/funding/learn/broader-impacts). "Look at this paper where there's a random high schooler on it" checks the box if a bit inappropriately. Incentives at work!

[SJC_Hacker]

That funny, at the time I left academia, I found almost the entire opposite attitude. The faculty openly lobbied against having high school and sometimes even undergrads in the lab. This was in fairly heavy duty experimental life sciences, where we did small animal research, so I guess it depends on field/university/department.

But in my lab at least, we did have the odd high schooler/undergrad, whom I had a part in mentoring. For the high schoolers, my PI at least privately admitted it was because their parents had "means" and it was about at least the potential for grant money.