[Spivak]

Preface: I scored extremely well on the SAT/ACT so I would have every incentive to keep the status quo. I think these tests should stick around as an option for students to prove their aptitude but that the requirement for them should be dropped.

Being smart does help you on these tests but not as much as you would hope. The things that meaningfully affect your score are studying for them and learning the specific material on the test. So there are two ways to do well: you study your ass off with test-prep materials or you go to one of the "good" high schools that tailor their entire curriculum to the ACT/SAT, AP and IB tests.

tl;dr these exasm only test how good you are at school and leave very little room to prove your aptitude in areas that aren't the primary subjects in school. CS being one area that until very recently was completely absent from all but the very best schools.

[dragonwriter]

> I scored extremely well on the SAT/ACT so I would have every incentive to keep the status quo.

I did too (well, the SAT; given that and that evrywhere I wanted to apply took the SAT, the ACT would have been superfluous), but...

> Being smart does help you on these tests but not as much as you would hope.

Its pretty much all being smart. Focussed study has some effect (and because small score differences at the high end make big competitive differences, can be worthwhile), bit don’t really do much.

Scores are quite tightly correlatee with IQ, which is why, e.g., MENSA accepts them in place of IQ tests.

> The things that meaningfully affect your score are studying for them and learning the specific material on the test.

Sure, those are the things in your control near the time of taking the test that affect your score, aside from “not getting wasted the morning of the test”. There’s not much you (or anyone else) can do after early childhood to significantly improve your probable IQ at the time you take the test, but that’s still the main outcome driver.

[Spivak]

> Focussed study has some effect

This statement requires some elaboration. Cramming for the SAT/ACT in the short time before the test has a small effect on your overall score. But if you go to a preparatory middle and high school where you will essentially spend 6-12th grade studying for the tests because the school designed the curriculum specifically to prepare you for them you will see a big difference.

Over the four years I spent in high school I was assigned as homework every. single. AP Calc I & II problem that had ever been published. Is it really that surprising that our class did well?

Another way of saying this.

* If you read voraciously as a child you will score extremely well on the reading section without really trying. You aren't smarter. You just spent a decade inadvertently studying.

* I was a math nerd as a child. I read math textbooks for fun. I was doing college-level math in 9th grade. And surprise to no one I scored almost perfect on the math section. But crucially, this didn't make me smarter than my peers. Literally anyone who had decided to spend their time at recess reading number theory textbooks would have done just as well.

[dragonwriter]

> Over the four years I spent in high school I was assigned as homework every. single. AP Calc I & II problem that had ever been published. Is it really that surprising that our class did well?

AP tests are different than the SAT/ACT. AP tests measure knowledge of something (the degree to which one can study the test independently of the notional subject matter may be debatable, sure), but SAT/ACT are very much proxy IQ tests. They may frequently be used together (but MIT is dropping SAT/ACT, not AP), but are not equivalent to each other.

> If you read voraciously as a child you will score extremely well on the reading section without really trying. You aren't smarter.

There is considerable, though as in most things intelligence-related, not conclusive, evidence that reading, especially early reading, does improve general intelligence.

> I read math textbooks for fun. I was doing college-level math in 9th grade. And surprise to no one I scored almost perfect on the math section. But crucially, this didn't make me smarter than my peers.

How do you know that you didn’t both do this in part because you were already smarter than your peers, and that doing it didn’t make you even smarter than you would otherwise have been?

> Literally anyone who had decided to spend their time at recess reading number theory textbooks would have done just as well.

Even if that was provably true, that doesn’t mean doing it didn’t make you smarter, since if it did it would presumably also have that effect on others who did it.

[floren]

Well, I scored 760 on the verbal because I read constantly and voraciously. I didn't "study" for it except in the sense that I'd spent the last decade with my face in a book on the bus, during lunch, as I walked from class to class, after I finished work in class, and so on.

Of course my aptitude at picking out synonyms didn't really indicate a damn thing about how well I could write an exam essay about the Reconstruction Era, so the fact that I got a good score without studying doesn't say that much for the test.

[Spivak]

I mean you can play this game with any subject. I scored a 5 on the CS AP tests without studying because I had been coding nonstop since I was 10.

You are proving my point pretty much exactly. You didn't score well because you were smart in some vague objective sense -- you scored well because you had essentially spent 10 years inadvertently studying.