I’m still digging out from under everything that happened the four days I was gone. Normal AI posting should resume tomorrow, likely with a report from the conference.
Is Our Children Learning
The declines in NAEP scores and other tests seem to be concentrated on the weakest students. The bottom of the distribution is falling through the floor and the rest is mostly holding up. This does not match anecdotal reports coming out of schools, which suggest widespread declines, although those are presumably untrustworthy and it’s traditional to think children are in general getting dumber. That doesn’t mean they aren’t but we need systematic evidence.
I mean, in addition to this meaning essentially college-level reading skills, it’s very obviously not true. The statement makes absolutely no sense, and is based at best on an Obvious Nonsense theoretical distribution of scores.
Oli (top comment): There are two private schools in the capital, Lusaka, American International School of Lusaka and International School of Lusaka.
Gerry Tan: If you *only* used SAT to admit to elite colleges, share of admits from top 1% income falls 15.8% → 9.9% and representation from <$200k rises by +8.8%, with no reduction in post-college outcomes.
It’s ‘holistic review’ and ‘ban SAT’ policy that allows the most wealthy and powerful to virtue signal while getting an edge.
Paper Abstract: The high-income admissions advantage at Ivy-Plus colleges is driven by three factors: (1) preferences for children of alumni, (2) weight placed on non-academic credentials, and (3) athletic recruitment.
…
The three factors that give children from high-income families an admissions advantage are uncorrelated or negatively correlated with post-college outcomes, whereas academic credentials such as SAT/ACT scores are highly predictive of post-college success.
Non-academic credentials, including most athletic slots (since most are not the popular and accessible sports like basketball or football), end up largely being manufactured by the wealthy by spending money and performing class. As commentator Naveen says, the main flaw in the SAT/ACT is that they run into big time ceiling effects, and can’t differentiate top 1% from top 0.1% or top 0.01%.
There is a very easy and obvious fix for that, which are academic credentials that do differentiate at that level. What, like it’s hard? Even if we can’t do better than this, and we have to do exactly zero innovation, we can give the students eyeing Ivy+ level colleges the GRE or the LSAT or the AIME. Check their AP scores. And so on.
Do Not Saturate Your Benchmarks
One main problem with SAT and ACT scores is that they max out. If you’re trying to get into Harvard, you have a 2400 or 36, but mostly so does everyone else. There is a simple solution to this, which is to extend the SAT and ACT so the top score is higher, and the toughest questions are a lot tougher, and no one ever gets a perfect score. This only impacts the very top directly, but it also means one can aspire to that elite and make it count, shifting incentives.
There are other standardized tests like AP exams, and in theory one could jump ahead and take a GRE test in 11th grade, but in practice these don’t do the job.
Some even start tiny non-profits, which is a complete waste of time. I think pressuring kids to volunteer for ‘community service’ or at nonprofits to graduate high school or sell themselves to colleges is analogous to indentured servitude and we should look upon it in horror. Legally mandated highly ineffective altruism mixed with child slavery. And this is if anything stupider and worse.
Marc Porter Magee: The kids know this is a meaningless exercise. Their parents know it too. Even the college admissions teams must know it. And yet it rolls forward, getting more intense each year.
Having started a nonprofit, I feel confident in saying incentivizing children to do it to gain an edge in a college application is a huge waste of everyone’s time and money.
It usually takes a year just to get your IRS determination letter. There is a ton of overhead involved, which isn’t worth it unless you are raising a ton of money over a long period of time.
How about start by volunteering somewhere you believe in?
Should kids be pushed to file with the IRS to take in $1,000 in donations to give sunscreen to 250 people? What exactly are we incentivizing here, the trappings of founding something?
This isn’t a question of kids having a summer job or volunteering. It’s telling kids that’s not good enough. They can’t just volunteer, they have to start a nonprofit. They can’t just have a summer job, they have to found a company.
I too have started a 501c(3) nonprofit, Balsa Research. I believe in the work we’re doing, and also it’s very good for someone like me to have a 501c(3) in reserve in case I need one in a hurry for something else. But damn if it wasn’t an insane amount of overhead, in terms of the money and time and work required to have it exist at all. Doing this as a high schooler is mainly a way to learn to hate government red tape.
I strongly agree that the Gaokao is a better way. It’s not first best, since it creates its own kind of all consuming zero sum competition, but it seems a lot more actually productive and less destructive.
Holistic Admissions Should Mostly Be Positive Selection
If you do holistic admissions, and people do something amazingly great, admit them.
Hannah Cairo solved a 40-year-old mystery about how functions behave, called the Mizohata-Takeuchi conjecture, which she proved false while everyone else was trying to prove it was true. She was taking graduate classes in mathematics at Berkeley.
So she skipped college and applied directly to graduate schools. Despite being obviously qualified, 6 out of 10 rejected her for not having a degree, and in 2 more cases she was admitted and then higher-ups in the administration overrode (!) the decision, leaving only University of Maryland and Johns Hopkins University.
We had a similar situation in my high school class, where we had two supreme math talents on the IMO team. One, Alex, wanted to skip directly to graduate school. Which he did, but at the cost of going to what I think was UPenn, whereas the other similarly talented math student, Michael Devlin, let college happen and in exchange got to go to Harvard. I was not as impressive and got neither of these choices.
Beware Stolen Valor
Articles tried to call Hannah Cairo a ‘high school student.’
No. A person of the age where we would normally imprison her in high school and teach her nothing relevant. Big difference.
Simon Sarris: Earlier this year, a 17-year-old *homeschooled* student She finished calc at 11. Imagine if she was kept in public schools. Or worse those Californian public schools that try to stop kids from learning math “too early” (no algebra in 8th grade).
Tracing Woodgrains: Calling her a high school student is stolen valor for high school. In a regular school, her talent would likely have been somewhat wasted. It is almost certain she would not have progressed at the same pace. This is not an iron law of schools. It’s just what we’ve chosen.
The Name Game
Having a last name late in the alphabet will cost you 0.6 points on a 100 point scale when your work is graded in school in alphabetical order, which it often will be. This is one of many later-name-gets-punished mechanisms, and they add up, for both first and last names, more than people realize.
Fair Weather College
Students who tour colleges are more likely to attend if there was good weather when they went on their tour. Some adjustment is wise, since it is some hard-to-fake evidence of weather patterns, but presumably they are adjusting too much.
I mean, yeah, high school students use their feel and vibes to pick colleges within ranges, because they are teenagers, picked colleges that are close picks for them and have little else to go on. This shouldn’t be surprising or worrisome? But it is fun.
Disability Accommodations Are Now Mostly A Scam
Richard Hanania: The rise of disability accommodations for LSAT takers. Went from less than 1% of test takers to 11% in a decade.
The Atlantic: The increase is driven by more young people getting diagnosed with conditions such as ADHD, anxiety, and depression, and by universities making the process of getting accommodations easier. The change has occurred disproportionately at the most prestigious and expensive institutions. At Brown and Harvard, more than 20 percent of undergraduates are registered as disabled. At Amherst, that figure is 34 percent. Not all of those students receive accommodations, but researchers told me that most do.
Unless and until someone gives a strong counterargument, I buy Hanania’s explanation that the 2008 amendment to the ADA was bonkers crazy and the natural result is that essentially anyone who makes an effort can now claim to be disabled. It is fortunate for us that the solving for the equilibrium on this has mostly been contained to academia and in particular to testing, where it is mostly harmless.
Does it ruin standardized testing? It ruins meaningful time limits, or anything which can be modified by a required ‘accommodation.’ That’s annoying, but I do not see it as fatal. At this point, it makes sense to redesign the tests so that time is not a key limiting factor. That costs everyone time, but what is the alternative? Fix the law? That doesn’t seem likely.
Harvard Has Some Grade Inflation
Grade inflation at the undergraduate level is vastly worse than anything in high school. The average GPA at Harvard before recent efforts to fight it was a 3.8 with 79% in the A range.
Noah Smith: One reason higher education in America is due for a massive shakeout is that undergrad grading has become a total joke.
Steve McGuire: Harvard faculty recognize that grade inflation has become absurd.
Rose Horowitch: During their final meeting of the spring 2024 semester, after an academic year marked by controversies, infighting, and the defenestration of the university president, Harvard’s faculty burst out laughing. As was tradition, the then-dean of Harvard College, Rakesh Khurana, had been providing updates on the graduating class. When he got to GPA, Khurana couldn’t help but chuckle at how ludicrously high it was: about 3.8 on average. The rest of the room soon joined in, according to a professor present at the meeting.
They were cracking up not simply because grades had gotten so high but because they knew just how little students were doing to earn them.
…
And yet, these students report being more stressed about school than ever. Without meaningful grades, the most ambitious students have no straightforward way to stand out. And when straight A’s are the norm, the prospect of getting even a single B can become terrifying. As a result, students are anxious, distracted, and hyper-focused on using extracurriculars to distinguish themselves in the eyes of future employers.
…
Without the threat of poor grades, students have largely stopped trying in their courses.
As you would expect, you put the kids that won the competition to get into Harvard into Harvard, take away the competition for grades, and you get other competitions stepping up to fill the void, except now no one is focusing on the classes.
The kids do vastly prefer this to the alternative of being actually graded. They fought very hard for the right to get there and not be actually graded.
Steve McGuire: Harvard reports that it is “failing to perform the key functions of grading.”
Its grading practices are “damaging the academic culture of the College.”
“Faculty newly arrived at Harvard are surprised at how leniently our courses are graded.”
Students say academics feel “fake.”
Grade inflation is out of control: 60% of grades are now A’s. That’s risen from 26% just 20 years ago.
The median GPA at graduation has risen from 3.29 in 1985 to 3.83 last year.
Students, for their part, were more sanguine. ‘I think the current grading system is working very well,’ one reported, ‘I haven’t heard many complaints from other students either.’”
I’m sure…
The report concludes that the grading system is failing because it doesn’t:
motivate students (they do enough to get an A and stop there)
communicate mastery of course materials (grades now only indicate whether work “meets expectations”)
distinguish the strongest students
It is ruining the undergraduate culture because students:
choose courses based on grade expectations rather than interest
are more stressed out as they try to distinguish themselves in other ways
say academics feel “fake” because there is no sense of achievement.
Well, sure, if academics don’t provide challenge, achievement or ability to outperform then, in a college of the nation’s top performers, they will (as we’ve discussed several times before) focus on other means of competition and differentiation.
It’s even worse than it looks because Harvard doesn’t allow the A+. One proposal is to allow a limited number of A+ grades, so at least there’s something to compete for.
The most incredible part of this report might be when Harvard acknowledges that its pedagogical crisis is partly due to coddling underprepared students:
“For the past decade or so, the College has been exhorting faculty to remember that some students arrive less prepared for college than others, that some are struggling with difficult family situations or other challenges, that many are struggling with imposter syndrome-and nearly all are suffering from stress.”
“Unsure how best to support their students, many have simply become more lenient. Requirements were relaxed, and grades were raised, particularly in the year of remote instruction. This leniency, while well-intentioned, has had pernicious effects.”
Harvard Crimson: She also connected the shift to efforts from the College to encourage instructors to be sensitive to students who enter Harvard with limited high school preparation.
“For the past decade or so, the College has been exhorting faculty to remember that some students arrive less prepared for college than others,” she wrote. “Unsure how best to support their students, many have simply become more lenient.”
There should be no need to coddle unprepared students at Harvard, because there should be no need to ever admit an unprepared student. Like, students who enter Harvard with limited high school preparation, what? I would simply not admit them.
Steve McGuire: These Harvard students…did not react well to the report on grade inflation:
“The whole entire day, I was crying. I skipped classes on Monday, and I was just sobbing in bed because I felt like I try so hard in my classes, and my grades aren’t even the best. It just felt soul-crushing.”
“What makes a Harvard student a Harvard student is their engagement in extracurriculars. Now we have to throw that all away and pursue just academics. I believe that attacks the very notion of what Harvard is.”
“I can’t reach my maximum level of enjoyment just learning the material because I’m so anxious about the midterm, so anxious about the papers, and because I know it’s so harshly graded. If that standard is raised even more, it’s unrealistic to assume that people will enjoy their classes.”
A student says harder grading “could take a serious toll on students’ mental health.”
“‘It makes me rethink my decision to come to the school,’ she said. ‘I killed myself all throughout high school to try and get into this school. I was looking forward to being fulfilled by my studies now, rather than being killed by them.’”
Harvard students are in revolt over this grade inflation report:
“Upvote to fire Amanda and anyone else who want [sic] to keep screwing our GPAs.”
“Protest happening tomorrow at Harvard Statue 5pm against deflation. Bring your books. Bring your syllabus. Bring your lighters.”
Timothy Taylor: The good news: Student learning takes off at Harvard (with especially great student learning during the pandemic!), as shown by rising percent of A grades in Harvard classes since 2005
Fighting Grade Inflation with GAAP Accounting
Thus, Harvard decided things had gotten rather out of hand, and tried a grade cap.
Alex Tabarrok says this makes them not a good solution, but I would compare this to the situation as it is now. Don’t tougher classes usually give out less As not more? There are some exceptions, but lots of dirt easy courses are mostly As and by default very hard courses are relatively stingy. So forcing this to line up is helpful on the margin.
If you want to solve the problem of true course difficulty, as in controlling for student ability and effort, then you need to control for that explicitly.
There actually is an important modification.
Scott Kominers: Harvard aims to curb grade inflation by capping A grades in each course at (20% of enrollment) + 4 and evaluating honors using average percentile rank. I’ve studied grading through the lens of incentive design.
The verdict: Harvard’s proposal fails the game theory exam.
The modification is that a smaller class can let that +4 go a long way. If you have only five students, then you have zero problems.
My solution is simple, which is that we scrap the letter grades entirely, and only offer percentiles. Except, as per tradition, we call them something else: Strength assessments.
As in, we know everyone at Harvard is super strong, so we assign them all a strength score of 18 in all subjects. Then the lowest percentile is 18(01), so everyone knows it is still impressive, and the best are 18(99) and then 18(00), the top of the human range. That’s grade inflation done right, and then everyone knows to adjust for difficulty.
I am, of course, largely kidding with that.
But also there’s an easy way to not be kidding, which is to offer a clear statistic to compare performance between different classes. Sports fans have done many similar things, and know how to adjust for strength of schedule and even moves between minor and major leagues, so we all know this is not so hard.
Thus, I propose Grade Average Above Par (GAAP) Accounting.
Par performance is defined as how a marginal admitted student would perform, and we use standard sports-style statistical methods to compare high school and college grades to form predictions of that score for every class, based on who took the class and how those students perform elsewhere.
Then you can extend this across schools, since you can compare their matching high school transcripts and standardized test scores, to get a universal U-GAAP score, also known as Real GPA (RGPA), which can then be comparable across schools, and can predict how a Michigan student would have done at Harvard, or on the other end at Ohio State.
You can say that if employers look at GPA rather than U-GAAP, it won’t matter. But as a school you have a lot of control over this, because everyone is lazy, and you can report the GAAP as GPA and RGPA, and only at most report in an obscure location the old raw GPA as NGPA (nominal GPA), or perhaps not at all.
After all, if you can’t calculate your own NGPA, you don’t deserve to benefit from it.
This is not how bell curves work. If over 50% of students get As, or even a lot more than that, then yes I understand this is Yale so the students would be getting As at other schools but these grades are a lot less informative or useful.
Is Our Children Learning?
This sounds about right.
leah: if anyone’s wondering what the undergrad classes are like these days
Tim Timbo: 10 people speaking in an undergraduate class is way over preforming and he’s complaining lol
j⧉nus: when i was in undergrad the STEM classes were mostly fine. but people in the humanities (e.g. philosophy) classes seemed extraordinarily dumb. i often didn’t read assignments or pay attention, but i was consistently the only kid able to carry discussions or answer questions. it was pretty awkward because i’d try to give others a chance so whenever the professors asked a question there would be like 15 seconds of uncomfortable silence and then i’d answer.
im actually not sure why this is. i went to a highly competitive university and the people there shouldnt have been so dumb at humanities but they were in every humanities class i took.
It’s in the top 20 US universities.
Biting All The Bullets
If students realize that under no circumstances will there be any consequences for their failure to do anything, why should they do anything?
Jes Bettis (via Bluesky): An issue we’re seeing at all levels of university is that many students are simply refusing to do anything. They aren’t reading the syllabus, aren’t following assignment guidelines, aren’t engaging with material, ignoring deadlines. And this might seem like old news, but it truly has ramped up.
We’re seeing possibly the worst attendance in history. Even in creative writing classes, where students generally want to participate and have their voice heard, they’re not showing up, ignoring guidelines and feedback, and either describing expectations as “unclear” or simply refusing to comply.
We keep adjusting our goalposts, trying to meet students where they’re at, but for some, that place is a kind of non-place, where they don’t complete any work or engage with the class, but still receive a high grade. Emails are ignored. They can’t meet a deadline or keep an appointment.
The other reasons to do the things we are asking students to do would be ‘because it is useful, or fun, or interesting,’ or otherwise valuable. The students are unconvinced.
If the students have realized that there are no consequences, and thus no incentives, and are thus providing no response, perhaps they have learned the most valuable lesson currently available to them? And perhaps it is the university that needs to learn.
Good Luck, Have Fun
Zack Kessel: Apparently, college has gotten so easy that professors no longer teach the difference between “insure” and “ensure.”
The actual central thing here is that everyone involved seems to have stopped pretending college is about learning or acquiring skills?
College is getting easier.
The trick is to ensure it’s still ‘stimulating and fun.’
Profit. For the college, presumably. Cause you pay them.
Steve McGuire: An NYU professor AI-proofed his assignments.
The students complained they were too hard and that “he was interfering with their ‘learning styles.’”
Tyler Austin Harper: The takeaway here should not be that students are lazy. 18 year olds aren’t the problem. The problem is that universities have failed to make a persuasive case to their “customers” that they are educational institutions rather than companies that dispense “I’m employable” stamps.
There is a lot of ugly finger pointing and circular-fire-squading on the AI issue — “It’s the professors’ fault! No it’s the students’ fault!” — all of which has the convenient effect of letting the “AI is a tool” admins off the hook for their complacency and abysmal leadership.
Zac Hill: I am more willing than Tyler to say that 18-22y/os (ie adults) are lazy, but I do think this broader point is 100% accurate. Colleges/universities’ conception of their mission is broadly out of step with their customers’ expectations.
Tyler Austin Harper: I agree that they’re often lazy in the way that 18 year olds can be, but the problem is that nothing about the structure, administration, or financial model that pervades higher ed leads students to conclude that learning is actually valued, so why should they themselves pretend?
Zac Hill: Yeah that’s well-put and I agree that is the root issue. The rejection of learning as an ordinal imperative pervades at every level.
From the OP, this is the key thing, put better than I’ve previously seen:
Clay Shirky: The assignment itself is a MacGuffin, with the shelf life of sour cream and an economic value that rounds to zero dollars. It is valuable only as a way to compel student effort and thought.
Thus, you have to convince the student they want to do this particular effort and thought, or you have to force them to do it. Or they won’t do it.
Earlier this semester, an NYU professor told me how he had AI-proofed his assignments, only to have the students complain that the work was too hard. When he told them those were standard assignments, just worded so current AI would fail to answer them, they said he was interfering with their “learning styles.”
He was indeed.
A student asked for an extension, on the grounds that ChatGPT was down the day the assignment was due.
Another said, about work on a problem set, “You’re asking me to go from point A to point B, why wouldn’t I use a car to get there?”
It’s exactly the right question to ask why you wouldn’t use a car, the obvious but enlightening answer is you’re doing it for the exercise not the destination, and this would be like being driven around at a track meet.
And another, when asked about their largely AI-written work, replied, “Everyone is doing it.”
Those are stories from a 15-minute conversation with a single professor.
As a student I essentially goal factored, with ‘get good grades’ and ‘learn useful things’ as distinct targets. Only sometimes did I care about both.
You can obviously get around AI issues by doing oral exams, if you want it enough. If you do enough oral exams along the way, as they do in law school, that seems like a good way to actually get students to learn whether or not you assign and grade homework per se. Even if you’re not called on, it’s not hard to imagine what would have happened if you had been.
I only had one (one!) class in college where the grade was based on in-class Socratic methods.
Whether *you* learned the material is apparent in whether you can answer questions in front of the entire class.
Homework itself is the bad paradigm.
The class was Financial Markets and Intermediaries with Professor Meir Kohn btw.
It was widely considered one of the brutally hardest classes in the economics department.
The material wasn’t hard. It was the only one that *properly verified you understood it* week by week by week.
The class was Financial Markets and Intermediaries with Professor Meir Kohn btw.
It was widely considered one of the brutally hardest classes in the economics department.
The material wasn’t hard. It was the only one that *properly verified you understood it* week by week by week.
Arbutus Tree: Homework as a method of evaluation is a bad paradigm. But for things like math, you’re going to need to practice, get in the reps. You could maaaybe get enough of that in a reverse classroom?
William Eden: No, you just fail the people who don’t do the reps on their own. You can suggest problem sets but not grade them.
Kamila: i talked to my 14 yr old brother about the current state of the american public education system.
>teachers are using AI to grade assignments. imagine getting “you’re absolutely right!” on the canvas rubric
>most AI tools are blocked but the entire student body has come together to think up ways to bypass it. vpn’s, sketch websites, jailbreaking their chromebooks. they’re putting in more effort to workaround AI bans to do their homework than they would just doing their homework
>reading levels + comprehension are at an all time low. his english teacher quit after teaching for 15 years because she crashed out about students being, quote, genuinely stupider
Isaac King: Homework shouldn’t exist anyway. 6 hours a day is plenty for well-designed education, and kids already used their parents and the internet to cheat since well before AI was a thing. This doesn’t seem like as big a problem to me as people are making it out to be.
Most College Students Fake Wokeness
What does not stop upon admission is the preferences kayfabe.
It is rather obvious that on college campuses there is a lot of preference falsification going on, and a lot of lying about beliefs, and a lot of lying about your beliefs to other people who actually share your beliefs but are also lying the same way you are.
I still had no idea things were this extreme.
Steve Stewart-Williams: 88% of students report pretending to hold more progressive views than they really do.
Forest Romm and Kevin Waldman: Between 2023 and 2025, we conducted 1,452 confidential interviews with undergraduates at Northwestern University and the University of Michigan. We were not studying politics — we were studying development. Our question was clinical, not political: “What happens to identity formation when belief is replaced by adherence to orthodoxy?”
We asked: Have you ever pretended to hold more progressive views than you truly endorse to succeed socially or academically? An astounding 88 percent said yes…
…
78 percent of students told us they self-censor on their beliefs surrounding gender identity; 72 percent on politics; 68 percent on family values. More than 80 percent said they had submitted classwork that misrepresented their views in order to align with professors. For many, this has become second nature — an instinct for academic and professional self-preservation.
To test the gap between expression and belief, we used gender discourse — a contentious topic both highly visible and ideologically loaded. In public, students echoed expected progressive narratives. In private, however, their views were more complex. Eighty-seven percent identified as exclusively heterosexual and supported a binary model of gender. Nine percent expressed partial openness to gender fluidity. Just seven percent embraced the idea of gender as a broad spectrum, and most of these belonged to activist circles.
Perhaps most telling: 77 percent said they disagreed with the idea that gender identity should override biological sex in such domains as sports, healthcare, or public data — but would never voice that disagreement aloud.
That is some Soviet-level, reign of terror level preference falsification. We have a question where a full 77% of students that:
Believe [~X].
But will if asked in public will refuse to say [~X].
And will admit, to themselves and in a confidential survey, to both of these things.
That only leaves 23% total that either believe in [X], or are willing to say [~X].
That doesn’t mean a preference cascade is coming to campus any time soon. There is no reason this situation cannot be a stable equilibrium. It also seems likely to profoundly screw up the kids involved to be constantly lying in these ways.
First you get to college via the wrong metrics, as previously discussed in #12. Then things get easier.
I’m still digging out from under everything that happened the four days I was gone. Normal AI posting should resume tomorrow, likely with a report from the conference.
Is Our Children Learning
The declines in NAEP scores and other tests seem to be concentrated on the weakest students. The bottom of the distribution is falling through the floor and the rest is mostly holding up. This does not match anecdotal reports coming out of schools, which suggest widespread declines, although those are presumably untrustworthy and it’s traditional to think children are in general getting dumber. That doesn’t mean they aren’t but we need systematic evidence.
It’s Bad, But It’s Not That Bad
This really would be far worse than I think if it was true.
I mean, in addition to this meaning essentially college-level reading skills, it’s very obviously not true. The statement makes absolutely no sense, and is based at best on an Obvious Nonsense theoretical distribution of scores.
Standardized Tests Help Disadvantaged Students
Your periodic reminder. See this study, also this one and this one. SAT scores and other objective measures fight income inequality in college admissions.
Holistic admissions make it worse. But then everyone already knew that. A paper from Raj Chetty, David Deming and John Friedman brings the math.
Non-academic credentials, including most athletic slots (since most are not the popular and accessible sports like basketball or football), end up largely being manufactured by the wealthy by spending money and performing class. As commentator Naveen says, the main flaw in the SAT/ACT is that they run into big time ceiling effects, and can’t differentiate top 1% from top 0.1% or top 0.01%.
There is a very easy and obvious fix for that, which are academic credentials that do differentiate at that level. What, like it’s hard? Even if we can’t do better than this, and we have to do exactly zero innovation, we can give the students eyeing Ivy+ level colleges the GRE or the LSAT or the AIME. Check their AP scores. And so on.
Do Not Saturate Your Benchmarks
One main problem with SAT and ACT scores is that they max out. If you’re trying to get into Harvard, you have a 2400 or 36, but mostly so does everyone else. There is a simple solution to this, which is to extend the SAT and ACT so the top score is higher, and the toughest questions are a lot tougher, and no one ever gets a perfect score. This only impacts the very top directly, but it also means one can aspire to that elite and make it count, shifting incentives.
There are other standardized tests like AP exams, and in theory one could jump ahead and take a GRE test in 11th grade, but in practice these don’t do the job.
My actual belief matches Paul Graham, that this is all on purpose, as part of the War on Education. We don’t want to give smarter or more knowledgeable students a big advantage, so we make it hard to see.
Holistic Admissions Turn Childhood Into Kayfabe
Holistic admissions creates incentives to look good by holistic standards, as teens scramble to optimize their every moment, including an increasing emphasis on eating everyone’s summers, to score every point they can and convey a ‘story’ to college admissions. That includes choosing that ‘story,’ including what you aim to study, by junior year of high school at the latest.
That’s all in addition to rising pressures to engage in active lying, literal live kayfabe performances and preference falsification, or else.
Some even start tiny non-profits, which is a complete waste of time. I think pressuring kids to volunteer for ‘community service’ or at nonprofits to graduate high school or sell themselves to colleges is analogous to indentured servitude and we should look upon it in horror. Legally mandated highly ineffective altruism mixed with child slavery. And this is if anything stupider and worse.
I too have started a 501c(3) nonprofit, Balsa Research. I believe in the work we’re doing, and also it’s very good for someone like me to have a 501c(3) in reserve in case I need one in a hurry for something else. But damn if it wasn’t an insane amount of overhead, in terms of the money and time and work required to have it exist at all. Doing this as a high schooler is mainly a way to learn to hate government red tape.
I strongly agree that the Gaokao is a better way. It’s not first best, since it creates its own kind of all consuming zero sum competition, but it seems a lot more actually productive and less destructive.
However, as noted in #12, you cannot simply fall back on SAT and GPA given the pressures that would put on GPA, even if you also throw in harder standardized tests.
Holistic Admissions Should Mostly Be Positive Selection
If you do holistic admissions, and people do something amazingly great, admit them.
Hannah Cairo solved a 40-year-old mystery about how functions behave, called the Mizohata-Takeuchi conjecture, which she proved false while everyone else was trying to prove it was true. She was taking graduate classes in mathematics at Berkeley.
So she skipped college and applied directly to graduate schools. Despite being obviously qualified, 6 out of 10 rejected her for not having a degree, and in 2 more cases she was admitted and then higher-ups in the administration overrode (!) the decision, leaving only University of Maryland and Johns Hopkins University.
We had a similar situation in my high school class, where we had two supreme math talents on the IMO team. One, Alex, wanted to skip directly to graduate school. Which he did, but at the cost of going to what I think was UPenn, whereas the other similarly talented math student, Michael Devlin, let college happen and in exchange got to go to Harvard. I was not as impressive and got neither of these choices.
Beware Stolen Valor
Articles tried to call Hannah Cairo a ‘high school student.’
No. A person of the age where we would normally imprison her in high school and teach her nothing relevant. Big difference.
The Name Game
Having a last name late in the alphabet will cost you 0.6 points on a 100 point scale when your work is graded in school in alphabetical order, which it often will be. This is one of many later-name-gets-punished mechanisms, and they add up, for both first and last names, more than people realize.
Fair Weather College
Students who tour colleges are more likely to attend if there was good weather when they went on their tour. Some adjustment is wise, since it is some hard-to-fake evidence of weather patterns, but presumably they are adjusting too much.
I mean, yeah, high school students use their feel and vibes to pick colleges within ranges, because they are teenagers, picked colleges that are close picks for them and have little else to go on. This shouldn’t be surprising or worrisome? But it is fun.
Disability Accommodations Are Now Mostly A Scam
Unless and until someone gives a strong counterargument, I buy Hanania’s explanation that the 2008 amendment to the ADA was bonkers crazy and the natural result is that essentially anyone who makes an effort can now claim to be disabled. It is fortunate for us that the solving for the equilibrium on this has mostly been contained to academia and in particular to testing, where it is mostly harmless.
Does it ruin standardized testing? It ruins meaningful time limits, or anything which can be modified by a required ‘accommodation.’ That’s annoying, but I do not see it as fatal. At this point, it makes sense to redesign the tests so that time is not a key limiting factor. That costs everyone time, but what is the alternative? Fix the law? That doesn’t seem likely.
Harvard Has Some Grade Inflation
Grade inflation at the undergraduate level is vastly worse than anything in high school. The average GPA at Harvard before recent efforts to fight it was a 3.8 with 79% in the A range.
As you would expect, you put the kids that won the competition to get into Harvard into Harvard, take away the competition for grades, and you get other competitions stepping up to fill the void, except now no one is focusing on the classes.
The kids do vastly prefer this to the alternative of being actually graded. They fought very hard for the right to get there and not be actually graded.
Well, sure, if academics don’t provide challenge, achievement or ability to outperform then, in a college of the nation’s top performers, they will (as we’ve discussed several times before) focus on other means of competition and differentiation.
It’s even worse than it looks because Harvard doesn’t allow the A+. One proposal is to allow a limited number of A+ grades, so at least there’s something to compete for.
There should be no need to coddle unprepared students at Harvard, because there should be no need to ever admit an unprepared student. Like, students who enter Harvard with limited high school preparation, what? I would simply not admit them.
Fighting Grade Inflation with GAAP Accounting
Thus, Harvard decided things had gotten rather out of hand, and tried a grade cap.
Grade caps, such as saying Harvard classes can only give 20% As, can solve grade inflation per se, but they mean that students get penalized for taking harder classes.
Alex Tabarrok says this makes them not a good solution, but I would compare this to the situation as it is now. Don’t tougher classes usually give out less As not more? There are some exceptions, but lots of dirt easy courses are mostly As and by default very hard courses are relatively stingy. So forcing this to line up is helpful on the margin.
If you want to solve the problem of true course difficulty, as in controlling for student ability and effort, then you need to control for that explicitly.
There actually is an important modification.
The modification is that a smaller class can let that +4 go a long way. If you have only five students, then you have zero problems.
My solution is simple, which is that we scrap the letter grades entirely, and only offer percentiles. Except, as per tradition, we call them something else: Strength assessments.
As in, we know everyone at Harvard is super strong, so we assign them all a strength score of 18 in all subjects. Then the lowest percentile is 18(01), so everyone knows it is still impressive, and the best are 18(99) and then 18(00), the top of the human range. That’s grade inflation done right, and then everyone knows to adjust for difficulty.
I am, of course, largely kidding with that.
But also there’s an easy way to not be kidding, which is to offer a clear statistic to compare performance between different classes. Sports fans have done many similar things, and know how to adjust for strength of schedule and even moves between minor and major leagues, so we all know this is not so hard.
Thus, I propose Grade Average Above Par (GAAP) Accounting.
Par performance is defined as how a marginal admitted student would perform, and we use standard sports-style statistical methods to compare high school and college grades to form predictions of that score for every class, based on who took the class and how those students perform elsewhere.
Then you can extend this across schools, since you can compare their matching high school transcripts and standardized test scores, to get a universal U-GAAP score, also known as Real GPA (RGPA), which can then be comparable across schools, and can predict how a Michigan student would have done at Harvard, or on the other end at Ohio State.
You can say that if employers look at GPA rather than U-GAAP, it won’t matter. But as a school you have a lot of control over this, because everyone is lazy, and you can report the GAAP as GPA and RGPA, and only at most report in an obscure location the old raw GPA as NGPA (nominal GPA), or perhaps not at all.
After all, if you can’t calculate your own NGPA, you don’t deserve to benefit from it.
Not Fighting Grade Inflation
Let’s check over at Yale.
This is not how bell curves work. If over 50% of students get As, or even a lot more than that, then yes I understand this is Yale so the students would be getting As at other schools but these grades are a lot less informative or useful.
Is Our Children Learning?
This sounds about right.
Biting All The Bullets
If students realize that under no circumstances will there be any consequences for their failure to do anything, why should they do anything?
The other reasons to do the things we are asking students to do would be ‘because it is useful, or fun, or interesting,’ or otherwise valuable. The students are unconvinced.
If the students have realized that there are no consequences, and thus no incentives, and are thus providing no response, perhaps they have learned the most valuable lesson currently available to them? And perhaps it is the university that needs to learn.
Good Luck, Have Fun
The actual central thing here is that everyone involved seems to have stopped pretending college is about learning or acquiring skills?
Cheaters Gonna Cheat Cheat Cheat Cheat Cheat
The students have always largely preferred the learning style of ‘not trying.’
From the OP, this is the key thing, put better than I’ve previously seen:
Thus, you have to convince the student they want to do this particular effort and thought, or you have to force them to do it. Or they won’t do it.
He was indeed.
It’s exactly the right question to ask why you wouldn’t use a car, the obvious but enlightening answer is you’re doing it for the exercise not the destination, and this would be like being driven around at a track meet.
As a student I essentially goal factored, with ‘get good grades’ and ‘learn useful things’ as distinct targets. Only sometimes did I care about both.
You can obviously get around AI issues by doing oral exams, if you want it enough. If you do enough oral exams along the way, as they do in law school, that seems like a good way to actually get students to learn whether or not you assign and grade homework per se. Even if you’re not called on, it’s not hard to imagine what would have happened if you had been.
Most College Students Fake Wokeness
What does not stop upon admission is the preferences kayfabe.
It is rather obvious that on college campuses there is a lot of preference falsification going on, and a lot of lying about beliefs, and a lot of lying about your beliefs to other people who actually share your beliefs but are also lying the same way you are.
I still had no idea things were this extreme.
That is some Soviet-level, reign of terror level preference falsification. We have a question where a full 77% of students that:
That only leaves 23% total that either believe in [X], or are willing to say [~X].
That doesn’t mean a preference cascade is coming to campus any time soon. There is no reason this situation cannot be a stable equilibrium. It also seems likely to profoundly screw up the kids involved to be constantly lying in these ways.