New news from Harvard, where there is a proposal with widespread support to limit A grades to about 20% of each course, gift link:
Here’s the full proposal: https://bpb-us-e1.wpmucdn.com/websites.harvard.edu/dist/e/139/files/2026/02/Grading-policy-proposal-260205.pdf
Just catching up on this thread.
ChatGPT could be totally wrong, but I did a quick check of the SAT scores from 1984 to now.
Back then a 1500 was achieved by between 1500 and 2000 takers from a pool of 1.5-2MM, so a 99.89 percentile.
From 2016-now, 1500 is 98th percentile, which doesn’t seem like a big jump, but of the average 1.7-2.2MM who took it, between 30-45K kids got above a 1500. You can do the math between 0.11 and 2.00 or just subtract from the above, but to me, that’s a big jump.
I don’t think its so crazy to think that more smarter kids are going to Harvard and doing A-worthy work, and personally think of grades as relative mastery versus relative to your peers. If you demonstrate that you know the material cold, I’m not sure why it matters what other people do. These are undergrads. But I’m from the 80’s cohort, so dumber.
Not being from the US, I’m curious: was there the same level of test prep and number of rewrites that seem to be common among those aiming over 1500 in the 80s as there is now?
The SAT was revised in 1995 and 2016 in ways that shifted the score scales.
1500 → 1995 revision → 1540-1590 → 2016 revision → 1560-1590
However, there is also the effect of test preparation, test takers doing more tries, super scores, and more of the high achieving students aiming for more selective colleges and universities.
Back when I took the SAT, I did less than 15 minutes of test preparation, which was doing the sample questions in the booklet that the sign-up form came in. If I believe what is said on these forums, that would be the extreme low end of test preparation today.
I also do not recall it being common to try more than twice, and lots of students tried only once in early 12th grade.
The specific grade percentages from first and last sample years in the document linked earlier are below. Note that the % making non-A/A- grades more than doubled over this 12 year period. It’s not just a difference from 40+ years ago in the 1980s.
2012-13: ~62% A/A-, ~33% B grades, ~6% C or lower
2024-25: ~85% A/A-, ~13% B grades, ~2% C or lower
I don’t agree that SAT scores are a good representation of what grade distribution is expected, but I’ll continue with that. In 2024-25, 16% received non A/A- grades, so the relevant consideration for how many should receive non A/A- grades is 16th percentile SAT score at Harvard, not the typical SAT score. 2024-25 was a test optional year, so there are a significant portion of students with lower SAT scores than test-required years. As such I’d expect 16th percentile SAT to be substantially lower than many previous years. It wasn’t all 1500+ SAT scoring kids in 2024-25.
That said, I don’t think the correct solution is the proposal in the document, which sounds like it limits A grades (not A-) to the following. Just looking at class size is not adequate. One needs to consider what portion of students are successful in the class, which depends on which students are participating in the class. One should not expect the same grade distribution for a remedial version of an intro class vs an upper level class composed of students within their major + grad students.
No, test prep companies were relatively new in the 80s. In person, mostly classes.
I think that at that point in time, that’s where the top private schools really had an edge – they really knew how to prep kids, not simply for college but for admission to college. Not huge numbers of kids.
So yes, the question you are asking is the right one - do higher test scores mean a stronger student body. Not necessarily.
As for taking it more than once, in my circles, nobody did that unless something went dreadfully wrong the first time. Nobody I knew took it a second time for a better score if they had a reasonable score the first time. With that said, while I had many classmates who went off to selective schools, getting to the main campus of our state flagship was considered a major win.
Back in the late 1960’s at least where I lived we had no test prep at all. We just walked in off the street and took the SAT with no clue what it was going to be like and no preparation at all, other than breakfast and a shower. We took the SAT once and that was it.
This would not apply to most people on this web site: However, for those of us who were from outside the US, there was at least some vague belief that what was taught in English courses in the US and what was taught in English courses outside the US (such as in Canada) was not quite the same. As such, among the students in my high school who were superb at English, French, Latin, and anything else related to languages, it was still very rare to get as high as a 700 on the English part of the regular SAT. One single student in my high school got a 710, and he was probably the number 1 strongest student in the high school particularly in languages. In contrast, our math and science preparation seemed to cover what was on the SAT quite well.
Also, as @ucbalumnus has correctly pointed out, the scoring on the SAT has been revised to make it easier to get a higher score.
As such I do not know if we can conclude much by looking at SAT scores over different decades.
That was the case back in the day - everyone knew that the hardest thing about it was getting in. To a greater or lesser degree the same can probably be said about most of the Ivies (but not MIT).
I remember the SAT Verbal (English) section of decades ago being mostly English vocabulary, presumably as a proxy for how much reading one did (of course, what test prep that existed at the time often had lists of supposed “SAT vocabulary words” to memorize). Was it the case when you took the SAT?
I would think that someone who is good at English, French, and Latin would do well in English vocabulary tests, given the origins of many English words from French. But maybe they picked more obscure words that were not French-derived.
My problems with the English part of the SAT were very much vocabulary related.
It’s comical to think that all those As are legit. It’s also comical to think that a high SAT score means you are “smart”.
You think a dumb kid could get a 1500 on the SAT?
I think the SAT measures a very specific kind of capacity. It doesn’t mean you have analytical skills or good writing skills. It doesn’t mean you can quickly read and digest complex ideas. It’s am imperfect system. I work with medical students who have scored in the top 10% on the MCAT. It’s astounding what they are not good at.
Are those students therefore not smart? In what population of anything are there members who are good at everything?
We’re getting off the point. In a more competitive pool of students its not crazy to think that the students may actually be better than back in the day. That ingredient shouldn’t be lost in the salad of things leading to more A’s.
I have family who attended for grad school and it was common to joke that the “hardest thing about Harvard is getting in.”
In 1987, when I graduated HS, if you took the SAT more than once, i believe they averaged your scores. So one of the reasons there were fewer retakes if you did decently well was the risk that you would score more poorly and actually end up with a lower score overall. It was a gamble. Also, there wasn’t such abundant test prep…Kaplan was king.