Every teacher knows the feeling, when you’ve lost the middle of the room — that sinking sensation when half your students are raring to go and the other half are hopelessly lost, and nothing you do is going to reach both halves at once. It’s like bending over and hearing your pants go rriiip.
It’s what happens when the middle empties out.
By the analysis of U.C. Berkeley math professors, over three years (2021–23), the percentage of mathematically well-prepared students in Calculus I at U.C. Berkeley has dropped from 43 percent to 26 percent, while the percentage of severely underprepared students has risen from 26 percent to 37 percent — a widening split between the two extremes with less and less fabric between them. Students in the least-prepared tier had a 46 percent chance of not passing the class. A virtual coin-toss.
It’s the widening, U-shaped graph that U.C. Berkeley math faculty have been seeing and sounding the alarm about for several years now.
This is systemic strain, not just one isolated bad semester.
Ever since the U.C. system dropped the SAT and ACT as a required part of the application package, professors have had to develop workarounds for the fact that the admissions process — now using only transcripts and A-G course titles — no longer gives them comparable and reliable information about the mathematical preparedness of enrolled students. While a standardized test score means the same thing no matter which high school its taker attended, a GPA definitely does not. U.C. San Diego’s own math department tested this directly and found that the probability of all California high schools grading to an identical standard would be less than 0.01percent.
So the system started adapting.
At UCSD, admissions staff created a patch for the gap by building something they called the Math Index: a statistical model that combs through an applicant’s transcript — grades in basic math courses, how many advanced courses they took, what topics those courses covered, even which high school they attended — and spits out a single number predicting whether that student will need remedial math once they arrive.
That last variable is the tell. The model’s own designers had to code in “which high school” as a factor, because they’d already proven a grade at one school doesn’t mean what it means at another. Then they built — by hand, out of statistics — a rough replica of what a test score used to deliver automatically.
Other campuses have leaned harder on a different patch: comprehensive review, the practice of asking an admissions reader to weigh not just a student’s numbers but the “context” behind them — under-resourced schools, thin course offerings, family circumstance. It’s a well-intentioned instinct for shaping a cohort. But it’s also, at bottom, asking a human being to manually guarantee the comparability of mathematical preparedness a test score used to supply automatically.
Layer a few more patches on top — a diagnostic test given only after enrollment, a guaranteed seat at a less-selective campus for students admitted nowhere else in the system, a remedial course redesigned twice in two years to remediate skills gaps as far back as elementary and middle school math — and you get an entire system-wide admissions apparatus quietly reverse-engineered around a hole everybody can see but nobody’s closed.
U.C.’s own faculty knew this was coming. In 2020, a U.C. Academic Senate task force spent a year studying exactly this question and found, in its own words, that “perhaps counterintuitively … test scores were better predictors of outcomes for underrepresented groups than for majority groups.” Their recommendation: keep the tests.
“Obscuring these preparation gaps does not provide access; it provides a near-guaranteed introduction to failure for our most vulnerable students.”
— U.C. Berkeley math professors
Months later, a lawsuit over pandemic-era testing-center access — disabled applicants couldn’t reach accommodated testing sites during Covid — forced U.C.’s hand faster and further than its own task force had proposed. The settlement made all undergraduate admissions test-blind. The task force’s own findings never even got a Regents vote.
In fairness, that lawsuit responded to a genuine problem of mandated access. But five years later, U.C. finally moved to revisit the question. In June, the system announced it would formally study whether to bring standardized testing back, with a faculty working group set to report findings by mid-2027. Then, in mid-July — just days before the Board of Regents was to convene — that same faculty board abruptly voted to shelve its own review. The chair called it “revising its timeline.” The Board of Regents chair now says the study will wrap up “no later than” June 2027 – meaning Fall 2027 applications would still not include SAT/ACT scores.
This is the pattern, not an exception to it. A body with the standing to act on its own evidence keeps finding reasons to defer the decision.
Consider how two different U.C. campuses, looking at nearly identical evidence, chose to say this out loud.
U.C. San Diego’s own admissions workgroup spent a year building the Math Index, then interviewing the tutors who work with its least-prepared students. One tutor estimated that 90 percent of the Math 2 students she worked with couldn’t solve a parallel problem on their own, even right after being shown how. Another pointed to a structural problem underneath the numbers: some high schools were teaching math as “life skills” — calculators, the internet, prescribed formulas — without ever building what he called “mathematical thinking.” Faced with its own data, the workgroup’s formal recommendation was that “our representative on the Board of Admissions and Relations with Schools (BOARS) should advocate for a systemwide reexamination of the possible return to standardized testing” — a rather gentle recommendation the report itself backs not with its own devastating numbers, but with a citation to MIT, which reinstated the SAT in 2022.
U.C. Berkeley’s mathematicians, looking at their own Calculus I data over the same three years, didn’t hedge. In the data appendix to the two open letters – together signed by more than 3,000 U.C. faculty – the five STEM authors wrote: “The University cannot responsibly continue to ignore these foundational gaps. Restoring standardized data points is necessary to properly align academic support systems with student needs and ensure a viable path to success for everyone admitted to a STEM major.” Obscuring these preparation gaps, they added, “does not provide access; it provides a near-guaranteed introduction to failure for our most vulnerable students.”
Same university system. Same missing signal. Same three-year window. One faculty group asked, gently, that someone else look into it. The other told the university, in its own blunt terms, that it was failing its most vulnerable students by refusing to.
The difference isn’t the data. It’s who was writing, and how close they were sitting to the people with the authority to act on what they found.
One of the UCSD tutors interviewed for that same report had a student who’d never failed a single high school math class — not because he understood the material, but because, as he told her, his coach “had a special agreement with the teachers.” He arrived at U.C. San Diego as a first-year student whose grades indicated readiness to do advanced college-level math. But that wasn’t the case.
That’s what the 46 percent is made of. Not lazy students, not bad teachers, not one villain — just a system that has lost its one common yardstick for meaningfully assessing mathematical preparedness and that now asks everyone downstream to improvise a replacement out of grades that were never built to give an apples-to-apples comparison.
The U.C. system didn’t set out to lose the middle of the room. But something that tries to fit everyone stops fitting anyone precisely, and the system has been feeling the strain for years now, quietly, one stitch at a time — until the seam gives out, and there’s no mistaking that sound.
