Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• Why the University of California built an entire public dashboard to correct a number its own campuses had been publishing for years

• How wide the gap runs at schools that admit by major, with figures from Washington, California, and Georgia Tech’s College of Computing

• Why this is a structural feature of how these programs are built, not a one-year anomaly that corrects itself

• Why the most selective computer science programs are staying tight even as national enrollment in the major is falling

For paid subscribers:

• A framework for sorting a client’s list into real tiers by major-specific admit rate instead of by a school’s overall reputation

• When applying undeclared and switching into the major later is a real path and when it is a trap, with the current data on both

• What a capacity-constrained major’s own admissions rubric says it is screening for, and how to build a senior-year schedule and application around it

Why UC built a dashboard to correct its own number

In fall 2025, the University of California published something unusual: a public data dashboard built specifically to correct a figure the university’s own campuses had been putting out for years. Major and discipline do not move the odds for most first-year UC applicants, the system said plainly in announcing the tool, but for a small number of high-demand programs they are close to the only factor that matters. Until that dashboard existed, UC only published the overall admit rate for each campus.

The numbers explain why UC bothered. At Berkeley, the overall admit rate for the fall 2025 entering class was 11%. Applicants to computer science were admitted at 6%. Applicants to arts and humanities programs were admitted at 25%, a more than four-fold spread inside one campus, one cycle, behind one published acceptance rate that most families read and stop there. At UC Santa Barbara, the computer science admit rate itself moved from 17% in 2024 to 34% in 2025, proof that these gaps are not fixed. They shift with applicant volume and available capacity, sometimes by a lot, in a single year.

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

How wide the gap runs elsewhere

The pattern is not a California peculiarity. The University of Washington publishes its own three-year average admit rates directly on its admissions site: 47% for Washington residents and 39% for non-residents, university-wide. Its direct-to-major pathway into the Paul G. Allen School of Computer Science and Engineering, the primary way most applicants enter the CS or computer engineering programs straight out of high school, runs 30% for residents and 2% for non-residents over that same three-year window. For an out-of-state applicant, the university-wide rate overstates the real odds by roughly twenty-fold. For the fall 2025 entering class specifically, the Allen School’s own reporting put the resident figure even higher, at 37%, while non-resident admits ran close to 5%, a gap that moves year to year but never closes.

Georgia Tech shows the same structure from the institution’s side. The College of Computing’s own account of its most recent admissions cycle put the figure plainly: 1,547 students admitted to study computer science or computational media out of 13,711 who applied, an 11% admit rate. That is the same cycle in which Georgia Tech’s institute-wide admit rate ran 30% for Georgia residents and 9% for everyone else. A Georgia resident reading only the institute-wide number is working from a figure roughly three times more generous than what the College of Computing reports for its own applicant pool.

Why this doesn’t correct itself

The reason this doesn’t correct itself is capacity, not admissions strategy. When the Allen School moved to direct-to-major admission in 2018, the university’s own announcement was explicit about why: interest in computer science had grown well beyond what the program could physically support, and the old system, in which any admitted UW student could aim for CS but only some would make it in competitively after two years of prerequisite coursework, was producing real strain on students and families who had already built a college plan around a major they might not get. Direct-to-major did not solve the capacity problem. It moved the rejection earlier, to the application, where a family can plan around it instead of discovering it midway through a degree.

Every school in this piece describes the same constraint in different language: faculty who can only teach so many sections, lab and studio space that does not scale with applicant volume, staffing ratios that cap enrollment regardless of demand. None of that is solved by a stronger essay or a higher test score. It is solved, if it is solved at all, by picking programs where the applicant math is survivable, and by knowing which of the workarounds families hear about from other parents are real.

The number moving in the opposite direction

Here is the part that makes this genuinely hard for a family to reason through on their own. Nationally, computer science is losing students, not gaining them. The National Student Clearinghouse Research Center’s most recent enrollment report found undergraduate computer science majors at four-year institutions fell 8.4% this spring, a drop of about 53,000 students, following an 8.1% decline the previous fall. Total undergraduate enrollment grew over the same period. Computer science is one of the only major fields moving against the national trend. “Computer science stands out for its enrollment declines”, said Matthew Holsapple, the Clearinghouse’s senior director of research, in the organization’s own report on the numbers.

A family hears that computer science is cooling nationally and reasonably assumes admission is getting easier. Sometimes it is, as UC Santa Barbara’s jump from 17% to 34% shows. At the programs with the deepest brand recognition and the tightest space, the softening has not shown up yet, because the applicant pool at Berkeley or the Allen School is not the pool the Clearinghouse is describing nationally. A small number of programs are absorbing outsized demand from a shrinking but still enormous pool of interested students. Both things are true at once, and a list built on either one alone, computer science is oversaturated so aim high everywhere, or computer science is cooling so admission is getting easier everywhere, will misprice a client’s real odds at the specific schools on their list.

Which means the list-building question is not whether to include a capacity-constrained major on a client’s applications. It is how to build that list once the admit rate a family sees on a school’s homepage is not the one that will decide the outcome.

Knowing the gap exists is one thing. Knowing what to do about it for a specific client’s list is another. That’s what the paid section covers: a three-tier framework for sorting a list by real, major-specific odds instead of a school’s overall reputation, a school-by-school verdict on when applying undeclared and switching in later is a real strategy versus a trap, backed by UW’s own published numbers on both, and how to read a capacity-constrained program’s admissions rubric to build a senior-year schedule around what it is screening for. If you’re reading on a free subscription, that’s the part built to change how you build a list this fall.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What the Department of Education’s accreditation rule changes, from how new accreditors get recognized to how transfer credit gets evaluated

• Why a negotiation that started in April compressed into a rule now open for comment through September 21, with a final version targeted for November

• How the fight already underway over law school accreditation shows what enforcement under the new rule looks like before it is even final

• Where support and opposition split, and where the fight is a legal one rather than a partisan one

For paid subscribers:

• A framework for identifying which schools on a client’s list carry real accreditor exposure, and which ones do not

• What the new 45-day credit transfer disclosure requirement means for the class applying this fall, since it takes effect the same month they enroll

• Why the legal uncertainty hanging over parts of this rule is a reason to build this diligence into permanent practice, not a reason to wait for the rule to survive court

• How to sequence this into intake differently for rising seniors versus rising juniors, and how to raise this kind of diligence with a paying family without sounding alarmist

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

The rule arrived days before its first real test

On August 21, Department of Education staff released an 86-page report recommending that the government strip the American Bar Association of its authority to accredit the nation’s law schools, citing nearly 30 alleged violations. The ABA has held that authority since 1952. The report landed one day after the Department published its Notice of Proposed Rulemaking on accreditation, and it is not a final decision: a federal advisory panel, the National Advisory Committee on Institutional Quality and Integrity, is scheduled to review the recommendation on September 23 and 24, Under Secretary Nicholas Kent has the actual authority to decide, and the ABA can appeal.

The rulemaking traces back to an April 2025 executive order directing the Department to overhaul how accreditors are recognized and what they are permitted to evaluate. A negotiating committee, the Accreditation, Innovation, and Modernization committee, met for two weeklong sessions in April and May of this year. On August 20, the Department published the formal proposal in the Federal Register, opening a public comment period that closes September 21. If the Department finalizes the rule by its November 1 target, it takes effect July 1, 2027, the same month the students applying to college this fall will be moving into their dorms.

One detail from the negotiation is worth noting for what it signals about how the Department is treating this process. The voting seat representing the Department on the negotiating committee, a role historically filled by a nonpartisan career staffer, went instead to a political appointee, Jeffrey Andrade, this cycle.

What the proposal changes

Four changes account for most of what practitioners need to track.

1. The first change loosens who can become an accreditor. New agencies currently must operate for two years and gather letters of support from institutions before applying for federal recognition. The proposal removes both requirements. The Department has already backed the shift with money: in January, accreditation reform was one of four priorities in a $169 million round of federal grants, with part of that funding earmarked for a handful of emerging organizations trying to become new accreditors. Since 1999, exactly four agencies have earned that recognition.

2. The second restructures who sits on an accreditor’s decision-making body. Under the proposal, a board member cannot vote on any standard that affects an institution or program where that member serves as an officer, director, or employee. Since most institutional and programmatic accreditors are built around peer review, meaning their boards are staffed substantially by administrators and faculty from the very institutions they oversee, this would force many existing boards to reconstitute themselves around volunteers with no institutional tie. The independence requirement goes further for accreditors connected to a trade association: no shared office space, no shared staff, no budget support. The ABA situation is the clearest early preview of what this looks like once enforced.

3. The third replaces vague outcome language with numeric standards. Accreditors must set minimum expectations for completion rates, licensing exam pass rates, and return on tuition investment, evaluated without reference to race, ethnicity, or sex, and must weigh intellectual diversity alongside the traditional measures of faculty qualification.

4. The fourth touches every family building a list right now. Institutions must presume that general education and elective credits from another accredited school will transfer, unless they can produce a detailed written rationale specific to the courses in question. Four-year colleges must provide a transfer credit evaluation at the time of acceptance, or within 45 days of receiving a transcript submitted afterward. Inside Higher Ed reported this year that a 2025 study found nearly four in ten American adults have tried to transfer college credit, and that 58% of them lost some of it in the process. A separate study of Texas institutions found acceptance rates for equivalent credits ranging from 52%-94% depending on which school was receiving them.

Where support and opposition split, and where the fight is a legal one

Kent has described the current accreditation system as functioning like the “central nervous system of higher education”, one where limited competition among accreditors has removed the pressure institutions would otherwise feel to improve. That is the core case for deregulation: four new agencies recognized since 1999, and high switching costs for institutions stuck with an accreditor whose standards no longer fit their mission.

Robert Shireman, a longtime accreditation expert and a Democratic appointee on the Department’s own accreditation advisory committee, called the draft regulations “a cluster bomb being dropped on American higher education”. He made the comment in April, as the first negotiating session opened, reacting to the department’s initial draft, Shireman and others generally accept that the old accreditation system needed reform. Their objection, echoed in formal statements from the Council for Higher Education Accreditation and the Council of Regional Accrediting Commissions, centers on peer review: the mechanism by which people who understand a specific kind of institution evaluate it has functioned as the main check against federal reach into what colleges are allowed to teach and how, and this proposal narrows that check in the name of removing bureaucracy.

Larry Schall, president of the New England Commission of Higher Education, made a more measured version of the same point in May, before the final language was public: “it’s very hard to say that it’s collectively reducing burden”. Wherever the rule removes a requirement, it adds a new reporting obligation elsewhere. The credit transfer rule is the clearest example: broadly popular, and expensive to carry out, particularly for the community colleges and small liberal arts institutions least equipped to absorb new documentation requirements without added registrar staff.

Underneath both positions sits a genuine legal question. The Higher Education Act bars the Secretary from regulating the substance of what standards an accreditor applies, only that the ten statutory categories, from student achievement to loan default rates, get evaluated somehow. Whether the Department’s proposal describes those categories in more specific terms, which is legal, or effectively dictates the content of the standards themselves, which is not, is exactly the question litigation will need to resolve once the rule is final. That uncertainty does not pause on September 21. For a family weighing four years of tuition against a degree whose accreditor might be mid-dispute before that student graduates, it is a diligence question with a real dollar figure attached, and most college lists right now are being built without it.

Knowing which four provisions are in this rule is one thing, knowing whether any of them actually touch the schools sitting on your clients' lists right now is another. Paid subscribers get the three-filter framework for sorting real accreditor exposure from noise, a plain-English breakdown of what the 45-day transfer credit rule means for the specific students applying this fall, and a script for raising this with a family.

If you're building lists or sitting across from families this season, the free version tells you what changed. The paid version tells you what to do about it.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What Zachary Bleemer and Jesse Rothstein measured, using more than a century of institutional spending data, and the specific multiple by which the spending gap has grown since 1969

• How the gap gets financed, which is not primarily through government funding, and what that means for where a full-pay tuition check ends up

• Why this paper isn’t just a restatement of the wealth-premium research already covered in this newsletter, and how the two studies fit together once you look closely

For paid subscribers:

• The specific, free federal tool you can use to pull this exact number for any school on a client’s list, and what counts as a meaningful gap versus noise

• The one detail inside a public flagship that can make the spending gap almost disappear for a specific student, and how to check for it before writing a school off

• A framework for using this number alongside cost and acceptance rate in a real list-building conversation with a full-pay family

• Why the full-pay tuition check itself is one of the biggest funding sources behind the exact stratification this paper documents, and what that means for how you frame the value question

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What Bleemer and Rothstein measured

Zachary Bleemer, an economist at Princeton, and Jesse Rothstein, an economist at UC Berkeley, built something unusual: a dataset spanning more than a century of American higher education spending, combining federal records including IPEDS with older sources like the Biennial Survey of Education and the College Board’s own historical archives. The paper is forthcoming in the Journal of Economic Perspectives, a peer-reviewed outlet.

Their central term is “the meritocratic consensus,” meaning the mostly unquestioned assumption that a college education’s quality should track a student’s own pre-college academic achievement. Students with stronger test scores and grades get sorted into better-funded institutions. Students with weaker records get sorted into institutions that spend less on them. The paper’s contribution isn’t pointing out that this happens. It’s measuring exactly how much, and how that amount has changed since the sorting first hardened into a system.

In 1969, the top quarter of four-year institutions by selectivity had per-student instructional spending about double that of the bottom quarter. By 1993, that ratio had grown to about 2.5 times. By 2019, it reached 3 times. The Ivy League’s spending advantage over flagship public universities specifically grew from 68% in 1929 to 242% in 2019. None of this is a story about a handful of schools pulling away from everyone else. It’s a structural widening across the entire hierarchy, public and private both.

How the gap gets financed

The paper traces exactly where the growing spending gap comes from, and the answer complicates a common assumption. Government funding, both direct appropriations and student aid programs, is distributed roughly evenly across more- and less-selective institutions. It is not the source of the widening gap.

What financed the roughly $10,000 per-student growth in the gap between top and bottom-quartile institutions from 1993 to 2019 breaks down specifically: about $7,900 came from relative increases in net tuition revenue at the higher-spending institutions, government research contracts contributed another $7,800, and private gifts and grants added at least $1,200, while relative government appropriations actually declined by about $2,100 over the same period. In plain terms, selective schools didn’t pull ahead because the government funded them more generously. They pulled ahead because they were able to charge more in net tuition and attract more research money and private giving, and they reinvested a meaningful share of that back into instruction.

That detail matters for how you frame this with a full-pay family, and we’ll come back to it in the paid section, because a full-pay tuition check is not a neutral data point in this story. It’s one of the specific mechanisms the paper identifies as driving the gap.

Why this isn’t just the wealth-premium piece again

If you read this newsletter’s earlier coverage of Chetty, Deming, and Friedman’s research on legacy and non-academic admissions preferences, this paper might sound like the same finding restated.

That earlier research measured how a student gets admitted: the specific mechanisms, legacy status, non-academic ratings, athletic recruiting, that create an admissions advantage for wealthy applicants at Ivy-Plus schools, holding test scores constant. This paper measures something downstream of admission entirely: how much a school spends on the students it enrolls, and how that spending gap has grown over 125 years regardless of exactly who gets in.

The 2 papers actually cross-reference each other, and reading them together sharpens both. Bleemer and Rothstein cite the Chetty research directly to make a point that cuts against the assumption that non-academic preferences are driving the sorting: they note that at one unnamed Ivy-Plus institution in Chetty’s own dataset, legacy preferences shift the class by only about 3 percentage points toward top 1% families beyond what standardized test scores alone would produce. Only about 7% of college students nationally are intercollegiate athletes of any kind, and just 2% of all college students are NCAA Division I athletes, the specific group eligible for the preferential treatment most people picture when they think of recruited athletes. Bleemer and Rothstein’s own conclusion is direct: athletics, legacy, and affirmative action all matter far less than test scores in shaping which students end up at which institutions. Test scores, not the mechanisms families spend the most energy worrying about, are what actually sorts most students onto radically different resource tiers.

Put the two papers together and the picture gets more useful, not more repetitive. The earlier research tells you which specific levers move a student’s odds of admission at the very top of the pyramid. This one tells you what a family is actually buying once admission happens, at every tier of that pyramid, not just the top of it.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• Where the “roughly 1 in 5 colleges rescind an offer” figure comes from, and how a 2026-dated article ended up quoting research from nearly 2 decades earlier without anyone catching it

• What NACAC’s own current reporting system says about whether this number has ever been re-checked, and why rescission isn’t on its list of topics published or planned

• Why the mechanics of transcript reporting matter more than the stat itself, including the specific reporting gap that can leave a family finding out about a revoked offer in July or August, after every other option is gone

For paid subscribers:

• What triggers a review, based on what’s independently verifiable rather than the aging survey data, including the specific categories beyond grades that account for a real share of rescissions

• A conversation script for a senior who’s already checked out, built around the real reporting mechanics rather than a scare number

• What to do if a warning letter arrives, and the specific window in which a response changes the outcome

• How to fold a grades check into the fall cadence you’re already running with every client, tied to the mid-year and final report dates that matter most

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

Where the number comes from

Ask most IECs where the “roughly 20% of colleges rescind at least one offer a year” figure comes from and the honest answer is usually NACAC, without a specific year attached. That’s worth pausing on, because the specific year matters here more than it usually does.

The figure traces back to NACAC’s Admission Trends Surveys from 2008 and 2009, and NACAC stopped asking the question after that. Education Week reported plainly, back in 2019, that NACAC “stopped surveying its members on this question in 2009”, with the last measured figure sitting at 22%, not the “roughly 20 percent” now in circulation. The breakdown behind that number is more specific than most citations bother to include: about 2 in 3 rescissions traced to a senior-year grade drop, roughly a third, 35%, cited disciplinary issues, and close to 3 in 10 involved falsified information on the application. Those specific proportions, not just the topline percentage, have circulated in blog posts, IEC training material, and financial planning sites ever since, almost always presented as current, and it’s worth noticing that an education trade publication was already calling this old news 7 years ago.

The article that prompted this issue is a useful, if accidental, demonstration of how that happens. It’s dated August 4, 2026. It quotes Melissa Clinedinst, NACAC’s associate director of research and the named author of the association’s last full State of College Admission report, saying plainly that “NACAC’s latest research found that roughly 20 percent of colleges rescind at least one offer in a given year.” Nothing about that sentence signals it’s old. But the article’s own footer discloses that it was originally published in July 2017 and has been updated since, and nothing in the piece suggests the quote itself was ever revisited. The most likely explanation is the simplest one: a real interview happened once, in 2017, referencing research that was already 8 or 9 years old at the time, and the quote has been carried forward through cosmetic updates ever since, gaining a fresh publish date without ever being asked again.

What NACAC’s own records confirm

NACAC discontinued its traditional annual State of College Admission report after the 2019 edition. In 2023, the association relaunched it as an interactive, web-based tool it describes as “reinvented”, publishing individual topics on a rolling basis rather than one comprehensive yearly document.

That new system’s own page is the clearest evidence available that this specific question hasn’t been re-asked. As of its most recent update, NACAC has published exactly 4 topics under the new format: school counselor caseloads, factors in the admission decision, selectivity and acceptance rates, and yield rates. 2 more are listed as upcoming: recruitment strategies, and early decision, early action, and waitlists. Rescission appears in neither list. There’s no indication anywhere in NACAC’s own published roadmap that this question is scheduled to be asked again.

None of this means the old number is necessarily wrong today. It means nobody currently citing it, including, it appears, NACAC’s own communications, can say that with confidence. A statistic this frequently repeated in professional advice ought to have a source newer than 2009.

Why this matters for IECs

Here’s what changes for how you talk to a family this fall: the precise percentage of colleges that rescind in a given year matters far less than how the reporting process actually works, because the mechanics are independently verifiable and the stat isn’t.

The College Board’s own guidance on this is specific and current. Colleges do not receive a student’s final senior-year transcript until June or July, well after decisions and deposits are behind a family. That gap is the actual risk, more than any national percentage. A senior whose grades slide in the spring may not hear anything from a college until after they’ve already turned down every other offer and made their plans, at exactly the point in the calendar when there’s the least room left to fix it. That single fact does more useful work in an intake conversation than a statistic from either 2008 or today would.

Knowing the stat is stale, doesn’t tell you what actually triggers a review, what to say to a checked-out senior, or what to do if a warning letter arrives. That’s the rest of this issue.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What’s genuinely still changeable in a junior’s course schedule this fall, and the general window most schools allow before a change requires an administrative override

• What the PSAT/NMSQT covers this year, including the real October 2026 testing window, and the one score that counts toward National Merit, which is not the one most families assume

• Why a junior-year preliminary list has to function completely differently than the list this same student will build 12 months from now

For paid subscribers:

• A framework for building that preliminary list wide enough to still mean something in the spring, without letting a 16-year-old anchor emotionally to 3 schools that will look nothing like their eventual list

• How to sequence a first campus visit so it calibrates a student’s sense of scale and setting instead of turning into a premature attachment to the wrong data point

• The session-by-session map for the full 60 days, built around the real October PSAT window, so a family has a real working document by the time scores come back

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What’s still open, and what already isn’t

Every fall, a version of the same conversation happens in an early junior-year intake: a parent asks whether their student should be taking a harder math class, or whether it’s too late to add a 5th AP for the year. The honest answer depends entirely on the calendar of the specific school, and that calendar is narrower than most families think.

Public and private high schools each set their own schedule-change policy, but the pattern across most of them is consistent enough to plan around. A student can usually swap a course, drop a level, or add one within the first 2 to 4 weeks of the semester without much friction, sometimes just a signature. After that window closes, most schools require a counselor override, and a growing number ask for a documented academic reason before they’ll approve one. This is worth confirming directly with each client’s specific school in the first session rather than assuming a national standard, since the exact cutoff varies by district and sometimes by department within the same school. What doesn’t vary is the shape of the problem: whatever change a family is considering for this semester needs to happen now, not in October.

This moment carries more weight for a rising junior than it will again until senior year itself compresses everything into a matter of weeks. Junior year course selection is the last real opportunity to signal an upward academic trajectory before senior year schedules get built off of it, and a schedule finalized without a second look in the first weeks of September is the schedule an admissions reader eventually sees on the transcript. If a student is capable of more rigor than the fall schedule currently reflects, this is the moment to act on it, not a general intention to revisit in the spring.

What the PSAT is for this year

The PSAT/NMSQT sits inside almost every junior-year plan, and it is also one of the most consistently misunderstood pieces of the whole calendar. The College Board’s own testing schedule puts the fall 2026 in-school testing window at October 1 through October 30, with many schools choosing to administer it on the designated Saturday date, October 17. Registration doesn’t run through the student directly. The school selects its date and handles the logistics, which means the first useful thing to confirm with any junior client this month is simply which day their school picked.

The confusion families bring into a session isn’t usually about the date. It’s about what the score actually does. A 10th grader can sit for the October PSAT, and plenty do, but that score functions purely as practice, a data point for the family and the school, with no bearing on anything beyond that. Only an 11th grader’s PSAT/NMSQT score counts toward entry into the National Merit Scholarship Program. A family that took the sophomore-year sitting seriously and is now treating the junior-year one as just another repeat of the same low-stakes test has the risk backwards. This is the one that counts, and it’s worth saying that plainly in the first session rather than assuming a family already understands the distinction.

The second misconception runs the opposite direction: families who assume the PSAT deserves the same weeks of dedicated prep that the SAT or ACT will get later. It doesn’t, for most students. Its real value is diagnostic. It gives a junior and their family an early, low-stakes read on how a student’s current skill level maps onto the SAT scale, which is genuinely useful for deciding, months in advance, whether the SAT or the ACT is likely to be the better fit and how much runway a student needs before a real testing plan begins. Treating it as a score to chase, rather than an instrument to read, spends effort in the wrong place at exactly the moment that effort is better spent on the course schedule question above.

Why a junior’s list can’t be a small senior list

The third piece of a junior’s first 60 days is the one most likely to go wrong, and it isn’t really about testing or scheduling at all. It’s the instinct, often coming from the parent rather than the student, to start building a real college list this early, the way a senior would.

A senior’s list is a decision document. A junior’s list, done well, is closer to a research instrument, and treating it like the former instead of the latter is where a lot of unnecessary attachment gets built in September that has to be quietly walked back the following spring. A 16-year-old who names 3 schools in October of junior year, based on a campus visit, a sibling’s experience, or a name they’ve simply heard the most, is naming preferences built on almost no real information about their own academic trajectory, their eventual test scores, or what they’ll actually want by the time an application is due. Locking a family’s attention onto those 3 names this early doesn’t protect the student. It just moves the disappointment, or the walk-back conversation, to a later and more emotionally loaded moment in the process.

What replaces that instinct, and how to structure a preliminary list that stays useful without doing this kind of premature narrowing, is where the rest of this issue picks up.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What a 2025 peer-reviewed study found when it asked admissions officers how they actually read course rigor at schools they don’t know well

• Why Phillips Exeter and Phillips Andover, two of the schools that helped build the AP program, no longer teach it, and what that decision looks like on a transcript today

• The other applicants who produce the identical paper signature for reasons that have nothing to do with resources, and why an unfamiliar reader can’t tell the difference without help

For paid subscribers:

• The exact information admissions officers told researchers they need from a school profile to read course rigor correctly, and how to write it for a school where the thin catalog is a policy, not a constraint

• The same framework applied to homeschool documentation and international or expatriate transcripts, where the identical misread happens for entirely different reasons

• What the counselor letter has to state directly, in language the profile can’t, and the audit that catches the gap before decisions arrive instead of after

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What a 2025 study found when it asked admissions officers how they actually read course rigor

A study published in the American Educational Research Journal in October 2025 examined a part of holistic admissions that has had surprisingly little direct research attention: how admissions representatives actually learn about the high school a student is applying from, and how that understanding shapes their reading of the student’s coursework. The researchers, led by University of Maryland professor Awilda Rodríguez, interviewed college admissions representatives directly and analyzed how they described building a picture of high-resourced versus low-resourced schools. Three findings stood out. Officers were less likely to understand what a low-resourced school actually offered in the first place. They applied a different, more skeptical standard when judging whether a student had made the most of a limited catalog compared to a full one. And underneath both patterns, the researchers found officers holding what they term deficit-oriented beliefs about students from schools with fewer course offerings, beliefs that shaped the read before any individual student’s specific record entered the picture.

That default isn’t irrational as a general pattern. A 2018 U.S. Government Accountability Office report examined public high schools nationally and confirmed what most admissions offices already assumed: schools serving more students in poverty, and smaller schools generally, do offer measurably fewer academic courses to prepare students for college. An admissions officer who has read thousands of files has genuinely seen that correlation hold, at schools where it happens to be true. What the new study identifies is what happens when that same pattern-matching gets applied to a school where the correlation breaks down entirely, and nothing in the file explains why.

The school that helped build the AP exam, then declined to teach it

Phillips Exeter Academy and Phillips Andover were both involved in the pilot study that became the modern Advanced Placement program, working alongside Lawrenceville School and a small group of universities in the early 1950s to design a way for high school students to take college-level coursework. Roughly half a century later, both schools walked away from it. Exeter’s own college profile, the document sent to every admissions office alongside a student’s transcript, states plainly that the school does not offer AP courses, because its faculty concluded its own curriculum already goes beyond what the AP exams test. Andover made a parallel decision in the same period, keeping only a small number of officially designated AP courses in math and science and declining to label the rest of its curriculum that way, for the same reason. Both schools remain among the most selective secondary schools in the country. Andover currently admits roughly 13% of applicants.

None of that reasoning shows up on a transcript. What shows up is a short list of course titles with no AP designation next to most of them, from a school with a smaller total enrollment than a typical public high school, read by an admissions officer who may never have handled a file from that specific school before. Set against the study’s own findings, that is close to the exact paper signature the researchers describe officers defaulting to a deficit read on. The difference is that nothing about it reflects a resource constraint. It reflects the opposite.

The same paper signature, for reasons that have nothing to do with money

Elite boarding schools that deliberately limit AP designations are one version of this. They are not the only one. A family that homeschools through high school, for reasons ranging from travel schedules to a preference for a more individualized curriculum, produces a transcript with no institutional course catalog behind it at all. A family posted abroad for work, sending a student through a small international or expatriate school with an unfamiliar grading scale and a total enrollment in the dozens rather than the hundreds, produces a school profile an admissions officer is unlikely to have seen before, describing a curriculum built around a different national standard entirely. In each case, the choice behind the thin or unfamiliar catalog has nothing to do with what the family could afford. In each case, the paper produced looks, at a glance, like the paper produced by a school that genuinely lacks the resources to offer more.

None of this gets fixed by writing a stronger essay. The read the study describes happens earlier than that, inside two documents the student never sees and rarely thinks about: the school profile and the counselor letter. Getting those two documents right, for a family whose thin catalog is a choice rather than a constraint, is a specific and learnable skill.

What closes that gap isn't vague advice to advocate harder for a student. It's a specific set of things a school profile and a counselor letter each have to say, stated the way an unfamiliar reader needs to hear them rather than left to be inferred. Below: the exact categories admissions officers told researchers they need to read a thin catalog correctly, that same framework rebuilt for homeschool and international transcripts, and the one sentence a counselor letter has to carry that no school profile can say on its own.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What Raj Chetty, David Deming, and John Friedman actually measured, using anonymized admissions files linked to tax records and test scores

• The 3 specific mechanisms that create the wealth premium at elite colleges, and the exact share of the gap each one explains

• Why the same premium disappears completely at flagship public universities, and what the researchers found when they checked whether any of it predicts a graduate’s actual success

For paid subscribers:

• How to walk a full-pay client through this data honestly in an intake conversation, without sounding cynical about a process you’re being paid to help them win

• Which of the 3 mechanisms is worth a family’s time and money, which one is already saturated among applicants with real hooks, and which one cannot be bought at any price

• The 1 piece of the gap that has nothing to do with any admissions office, and the specific, sourced case it gives you for treating a flagship public as a real contender rather than a backup

Upgrade to paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What the researchers actually measured

That gap is not news to anyone running an IEC practice. Every counselor who has sat across from two equally strong students, one who got in and one who didn’t, has a private theory about why. What’s new is that 3 economists just told you which parts of that theory are correct, which parts are not, and exactly how much each true part is worth.

Raj Chetty, David Deming, and John Friedman are not guessing from enrollment photos. Chetty runs Harvard’s Opportunity Insights lab, and for more than a decade his team has built its reputation by linking IRS records to institutional data that colleges would generally prefer to keep to themselves. For this study, the researchers obtained anonymized admissions files from several private colleges and matched them to federal income tax records and SAT or ACT scores. That match is what separates this paper from the familiar observation that wealthy kids end up at better schools. Most prior work compared students who were already enrolled and inferred the rest. This one can take 2 applicants with the identical test score, one from a family in the top 1% and one from a middle class family, and trace what actually happened to each application.

The schools in the study are what the authors call Ivy-Plus: the 8 Ivy League schools, plus Stanford, MIT, Duke, and Chicago, 12 institutions in total. The comparison group is flagship public universities. That comparison turns out to matter as much as anything else in the paper, and we’ll get to why in a moment.

The team first circulated an early version of this analysis in 2023. The peer-reviewed version, revised and finalized, ran in the February 2026 issue of the Quarterly Journal of Economics, and it is only now working its way into IEC-facing coverage as this fall’s list-building season gets underway, which is the reason it belongs in this issue now rather than 6 months ago.

The headline finding: a student from a top 1% family is more than 2 times as likely to attend an Ivy-Plus school than a middle class student with a comparable SAT or ACT score. Chetty and Friedman wrote in Time in January that if admissions offices treated top 1% applicants exactly the same way they treat everyone else with the same scores, the share of top 1% students at America’s most selective colleges would fall by close to half.

Here is where the paper gets genuinely useful rather than just confirming a suspicion. The researchers split that gap into 2 pieces. About two-thirds of it comes from top 1% applicants being admitted at a higher rate for the same test score. The remaining third comes from something else entirely: differences in how often equally qualified students at different income levels apply in the first place, and how often they enroll once admitted. Hold onto that second piece. It’s the part of this study that has nothing to do with what any admissions committee decides, and it’s the part your practice can actually move.

Three mechanisms, and exactly how much each one explains

For the two-thirds of the gap that comes from differential admission rates, the paper names the specific channels and puts a number on each one, which is rarer in this kind of research than you’d expect.

1. Legacy preference is the largest single factor, responsible for 46% of the admission-rate advantage. An applicant from a top 1% family whose parent attended the same Ivy-Plus school is 5 times more likely to be admitted than an applicant with a comparable test score, demographic profile, and admissions committee rating whose parent did not attend. The researchers tested whether this is really legacy status doing the work, or some unmeasured family quality riding along with it, by checking whether a legacy applicant at one Ivy-Plus school gets any advantage at a different Ivy-Plus school their parent did not attend. They don’t. The advantage is specific to the exact institution the parent attended, which is about as clean a signal as this kind of research produces.

2. Non-academic ratings account for 30%. This is the admissions committee’s assessment of extracurricular activity, leadership, and personal qualities, the part of the file that is, in practice, easier to build when a family has the time and money to fund it. A research placement, a specialized instrument, a summer spent founding a nonprofit rather than working a retail shift: these read as stronger non-academic credentials, and they correlate with income for reasons that have very little to do with a 17-year-old’s actual character.

3. Recruited athletics explains the remaining 24%. The sports that carry weight at these particular schools skew toward ones that require private coaching, club travel, and specialized equipment, which is the same underlying pattern showing up in a different part of the application. Recruited athletes make up 10% to 15% of the incoming class at these schools, according to Chetty and his co-author John Friedman.

Now the finding that should change how you talk about all 3 of these with a client. None of them predict what happens after college. In the paper’s data, legacy status, non-academic ratings, and athletic recruitment are unrelated to a graduate’s later earnings, unrelated to admission into an elite graduate program, and unrelated to landing a job at a top firm. Academic credentials, meaning test scores and grades, are the one part of the file that reliably does predict those outcomes. The researchers ran the counterfactual directly: remove legacy preference, non-academic weighting, and athletic recruiting from the admissions process, admit the same distribution of test scores the current class already has, and the share of Ivy-Plus students coming from the bottom 95% of family income rises by 8.8%. Nothing about academic rigor has to give for that shift to happen. The bar just moves to something that isn’t for sale.

Where the premium disappears, and what predicts success instead

None of this shows up at flagship public universities. The paper is explicit: children from high-income families have no admissions advantage at flagship publics once you control for test scores. Not a smaller advantage. Whatever is driving the tilt at Ivy-Plus schools, whether it’s legacy pipelines, an admissions committee’s weighting of the kind of resume only money can build, or the pull of expensive youth sports, it belongs to a small number of private institutions. It is not a general feature of how American higher education sorts students by wealth.

That distinction is worth sitting with before you build a fall list. The stakes are real. In the same piece, the authors point out that fewer than 1% of Americans attend one of the 12 Ivy-Plus schools, yet those schools produce over 13% of the top 0.1% of earners, a quarter of US Senators, half of all Rhodes Scholars, and three-quarters of the Supreme Court justices appointed in the last 50 years. Families are not wrong to care about this gap. Where most of them go wrong is assuming the whole gap runs through channels their money can influence.

Everything above this line is the finding. What you actually say to a client sitting across from you this week, the one who has spent 4 years and a great deal of money building exactly the kind of resume this paper says doesn’t predict anything, is a harder conversation. That’s what the rest of this issue is for.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• How more than 23,000 North Carolina seniors accepted 88,000 college offers this year without writing an essay or hearing back from an admissions office, and how it actually played out for two real students in one county

• How the same mechanism is scaling nationally through Common App and a handful of state programs, and why colleges are moving toward it heading into the 2026-27 cycle

For paid subscribers:

• What a direct admission offer actually replaces in an IEC’s workflow, and the specific parts of the job it doesn’t touch at all, including how colleges quietly decide, offer by offer, whether to attach real financial aid or just an admission letter

• The three ways an offer misleads a family if nobody reads the fine print, including the self-reported GPA problem that can leave an offer that looked settled unraveling in September

• A 5-step audit to run on every eligible client before letters start arriving in mid-September, including the sequencing decision that determines whether claiming an offer helps a student’s reach strategy or quietly undercuts it

• Why “this isn’t my clientele” is the wrong response even for practices that only run reach-heavy lists, and what this changes about what you should actually be charging for

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What actually happened in North Carolina this spring

This spring, two seniors at Union County Early College, Ava Dunham and Jordan Casiano Pineda, ended up on the receiving end of exactly that. A school counselor, Jeslyn McFadden, had introduced Dunham to a state program that had already matched her with colleges before she filled out a single application. Dunham logged in and found Western Carolina University, her top choice, sitting there as a direct admission offer. Casiano Pineda’s version of the same process, on his way to UNC Charlotte in the fall, took about 30 minutes and didn’t involve an essay. Neither of them is unusual. McFadden said she started encouraging students to log in once she learned the entire Class of 2026, statewide, had already been pre-accepted somewhere through the program.

That program is NC College Connect, and the mechanic behind it is worth understanding in plain terms, because it isn’t staying in North Carolina. If a student is on track to graduate, carries a weighted GPA of 2.8 or higher, and has completed a fourth math course, the state automatically matches them to a personalized list of colleges where they already meet the bar for admission. No essay. No test scores. No letters of recommendation. The student logs into a portal, sees the schools they’ve already been accepted to, and clicks a button to claim the spot. The offer is real, not promotional mail, and it isn’t binding, so a student can claim several and keep working the rest of their list at the same time.

The scale of it is the part that should get an IEC’s attention. Between August 2025 and January 2026, more than 23,000 North Carolina seniors claimed at least one offer, and together they claimed more than 88,000 of them, according to the state’s own Department of Public Instruction. That’s about 21% of the entire Class of 2026 accessing college admission through a portal instead of an application. 11 public universities, 29 private colleges, and all 58 of the state’s community colleges took part.

This is not just a North Carolina story

Common App runs a national version of the same idea, and it has grown fast. When Common App launched its direct admissions program for the 2025-26 cycle, more than 200 member colleges signed on, offering students the same basic deal: clear a college’s GPA and residency bar, and you get notified inside your Common App account that you’re already admitted, before submitting anything. Common App’s own materials for member colleges say most of those schools are returning for the 2026-27 cycle.

Common App is not the only national platform doing this, either. Niche runs a separate direct admissions program of its own, with more than 155 colleges participating this cycle. States are experimenting with the idea too, though not always the same way North Carolina does it. Texas has guaranteed automatic admission by class rank at its public universities since the 1990s, well before any of this existed, and Ohio launched its own guaranteed-admission program for top-ranked students in the 2025-26 cycle, with a newer, more direct version still in pilot design for 2026-27.

None of this is happening by accident. The Western Interstate Commission for Higher Education projects that the number of U.S. high school graduates peaked in 2025 and will fall 13% by 2041, and Carleton College economist Nathan Grawe separately projects the college-age population could shrink by around 15% between 2025 and 2029 alone. Colleges that used to count on a growing applicant pool every year are now competing for a smaller one, and automating the easiest admissions decisions is one of the more efficient ways to compete for it.

One more thing worth knowing

Unrelated to any of the above: the University of Michigan’s College of Literature, Science, and the Arts, its largest college, announced a pilot starting fall 2027 that replaces first-semester letter grades for incoming freshmen with a simple pass or no-credit mark on the external transcript. Instructors still assign real grades so students know where they stand, those grades just won’t count toward GPA or show up on the transcript colleges and employers eventually see. The college says the goal is easing the adjustment to college-level work in the first term. A handful of other schools, including Caltech, MIT, Swarthmore, and Wellesley, already do something similar for their own first-year students.

Whether any specific offer is good, bad, or right for a particular kid is a fit question, and fit questions are the easy part of this job. The harder question, the one that actually determines whether you’re useful to a family this fall, is what a direct admission offer changes about the conversation you have with them before Labor Day, and what happens if you’re still running that conversation like these offers don’t exist.

What we cover in the paid section. What a direct admission offer replaces in an IEC's workflow, and the specific parts of the job it never touches. Three ways an offer quietly misleads a family if nobody reads the fine print, including a self-reported GPA problem that most intake processes aren't catching. A five-step audit built to run on every eligible client before letters start arriving in September. And the honest case for why this matters even if your practice doesn't work with this kind of client directly, plus what it should change about how you price your own work.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• Why two students with the same strategy, the same list, and equally strong personal statements can be in completely different positions by the second week of September, and what the research on procrastination and task switching says about why this happens to organized students too

• The five specific ways execution breaks down inside a single application, each one invisible from the strategy conversation you already had in June

For paid subscribers:

• The content bank system that replaces essay-by-essay drafting with a small set of reusable material, and the session structure for building one with a client this month

• The single-source file that keeps a fifteen-application caseload from quietly contradicting itself across portals, and the specific places contradictions actually creep in

• The polish ceiling: a concrete diagnostic for knowing when a revision round is still improving an essay and when it has started erasing it

• The time-budget framework for allocating a student’s limited hours by which sections of the application actually move an admissions decision, not by which one feels most urgent this week

• The September check-in script built to surface all five failure patterns in one sitting, before any of them compound in October

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What the strategy conversation missed

Two students on the same caseload can arrive at August with the same list, the same testing plan, and personal statements you would call equally strong on the page. Six weeks later, one of them has three supplements roughed out, her recommenders confirmed since June, and a Sunday check-in that takes fifteen minutes because everything she has said about herself is consistent across every portal she has opened. The other is still deciding whether to swap two safety schools, has restarted her Why Cornell essay for the third time, and just noticed that the leadership title she gave her internship in the activities section does not match what she wrote about the same internship in a supplement drafted two weeks earlier.

Both students have a good strategy. Whatever opened the gap between them happened somewhere else, in the part of the process where a student sits down alone and actually executes against a plan that looked complete in June.

That gap is where an IEC’s real value is moving. A strong list and a strong personal statement are close to table stakes now, available from a competent counselor, a well-researched family, or a well-prompted AI tool. What is not commoditized, and what the second student above needed and did not get, is a system for executing a complex, multi-part, deadline-heavy project without losing coherence along the way.

Why this shows up first right now

The timing is not a coincidence. School is starting or about to. The Common App reopened for the new cycle on August 1, and the runway this newsletter has written about before, the ten weeks between June and mid-August built for finishing the personal statement and researching supplements, has just closed. A student walks into September carrying whatever state her application is actually in on the day that runway ran out, not the state the June strategy session assumed she would reach.

What shows up in that state falls into five patterns, and they are worth naming precisely because none of them are about a student’s intelligence or her commitment to the process.

1. Procrastination that survives good planning. This looks like a discipline problem and is usually something more structural. A 2007 meta-analysis in Psychological Bulletin, synthesizing decades of procrastination research, found the strongest and most consistent predictors were not fixed personality traits like neuroticism but task aversiveness and task delay: how unpleasant a task feels in the moment, and how far off its consequences seem until they suddenly do not. A 650-word personal statement due eleven weeks out satisfies both conditions cleanly. It is emotionally harder than most schoolwork a student has done, and its deadline sits far enough away to not feel real until it is close enough to feel unmanageable. A good college list does nothing to change either variable.

2. Treating each essay as an unrelated assignment. A student applying to eight schools with supplements is not writing eight independent short pieces. She is answering variations on a handful of underlying questions about herself, and a student who approaches each one as a fresh blank page pays a real cost for it. Sophie Leroy, the management professor who established interruptions as a distinct field of study within organizational behavior, found in a 2009 study that attention from one task does not fully disengage before a person moves to the next one, a lag she termed attention residue, and that the residue measurably drags down performance on the task that follows. A student rewriting her core material from scratch for every supplement is generating that drag a dozen times or more over a single application season.

3. Inconsistent information across applications. A student maintains her own record of activities, hours, titles, and dates across a dozen separate portals that do not talk to each other and do not flag it when the same fact gets stated two different ways. Nothing catches the discrepancy until a reader does, which is the worst possible place for it to surface.

4. Over-polishing until the essay stops sounding like the student who wrote it. This newsletter has covered what AI-smoothed prose does to an applicant pool’s collective voice. The same erosion happens through purely human channels. A personal statement that has passed through a fourth or fifth round of well-intentioned adult feedback often reads more competently and says less about the specific person who lived the experience, because each round tends to trade away the odd, particular detail for the safer, more generic one.

5. Spending hours on sections of the application that were never going to move the decision much, while the sections that actually carry weight get whatever time is left over.

Diagnosing these five patterns is one thing. Building a process that prevents them across an entire caseload, while application volume keeps climbing and the counselor’s role keeps shifting from admissions advisor toward the person who designs how the work actually gets done, is a different problem, and it is the one the rest of this issue is addresses.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What actually happened in ACT’s two nationwide technical failures over the past 13 months, confirmed through state education officials and ACT’s own on-the-record spokesperson, not test-prep chatter

• How ACT changed ownership twice in two years, first to a private equity firm in 2024, then to ETS on July 1, and what each organization has and hasn’t confirmed about what changes next

• Where testing requirements actually stand this cycle: a picture that’s both more mandatory and less universal than most coverage suggests

• A short update on a separate, fast-moving story: a federal judge just let an antitrust case over Early Decision proceed to discovery

For paid subscribers:

• The paper-versus-digital decision framework to use with clients this fall, built specifically from where the two documented failures actually occurred

• A three-bucket target-list audit your clients can run in one sitting to sort every school on their list by actual testing requirement status

• The specific verification habit worth adding to every testing timeline this cycle, and why “the school received the score” stopped being a safe assumption to skip

Upgrade to Paid

What happened, confirmed by the people who reported it

The first incident is documented directly by state officials and an independent education newsroom. On April 8, 2025, a malfunction on ACT’s servers delayed the start of the exam for thousands of students across Illinois and prevented others from moving between sections once they’d begun, according to Chalkbeat Chicago’s reporting at the time. About 11,000 students were affected. Illinois requires the ACT to satisfy a high school graduation requirement, so this wasn’t an optional test day for the students involved. State Superintendent Tony Sanders raised the issue directly with ACT’s president, according to Chalkbeat. Some schools resumed testing that day; others rescheduled. A spokesperson for ACT confirmed to Chalkbeat that a server malfunction caused the delay.

The second incident was broader by design, not by coincidence. On May 12, 2026, Illinois education officials again notified districts, this time that ACT was reissuing scores nationwide for every student who’d taken the test online during spring school-day administrations, not a regional problem. Districts in other states, including Ohio, got the same notice the following day. Juan Elizondo, ACT’s director of strategic communications, told Wisconsin’s WMTV that the root cause was in the process that makes scores comparable across different test forms and different testing conditions. “We caught that it didn’t work,” Elizondo said, and added that ACT had corrected the process and was rescoring the affected tests. Paper test-takers and Saturday national test-takers were not affected, only the online school-day group. ACT said no corrected score would come in lower than the original. ACT initially targeted June 2 for completion of the revised reporting process, although affected students began seeing scores restored in MyACT by May 22.

Both incidents landed inside a larger transition already underway. ACT shortened its exam and made the science section optional starting in fall 2025, part of a broader push toward digital administration, the same stretch in which both failures occurred. In 2024, ACT entered into a transaction with private-equity firm Nexus Capital Management that moved its operating business into a new for-profit public benefit corporation, while a separate Iowa nonprofit retained an investment in the company. Then, on June 30, 2026, ETS announced it was acquiring ACT from Nexus. Encoura is not part of that deal; it stays with Nexus. ETS is the nonprofit that administers the GRE and TOEFL and, until 2024, administered the College Board’s SAT under contract. ETS CEO Amit Sevak told Higher Ed Dive that ETS approached Nexus about the deal, not the reverse, and that ETS plans to keep ACT’s for-profit legal structure in place “for now,” language that leaves the door open to a different structure later. ACT CEO Steve Tapp, who took the role in October 2025, said joining ETS would let ACT “take what we’ve built and scale it within a broader vision for readiness.”

It’s worth being precise about what both organizations have actually confirmed versus what remains open. ETS’s own announcement and subsequent reporting from U.S. News state plainly that there are no changes to the ACT’s format, scoring, testing schedule, or score reporting as a result of the acquisition, and that ACT will run as a standalone organization for now. But Sevak also told Higher Ed Dive that leadership still needs to work out what comes after this year in more detail. Read together, the honest picture is: the announced position is that there will be no near-term changes to ACT's products or services during this admissions cycle; longer-term structural questions remain open.

There’s a piece of context that complicates that reassurance further. In January 2026, the Wall Street Journal reported that ETS was exploring a sale of its own two flagship exams, the GRE and TOEFL, in a deal valued at roughly $500 million, a report that Inside Higher Ed and Forbes both independently confirmed. Public reporting has not established whether that process remains active. Selling off two signature products while acquiring a major competitor’s exam isn’t automatically contradictory, but it undercuts any assumption that a nonprofit acquirer means more stability for ACT than a private equity owner would have.

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

Where testing requirements actually stand this cycle

Timing matters because of what’s happening on the admissions side. We’ve tracked the Ivy testing reversals across several issues this year. What’s new since the last one: Columbia announced in June that it will follow Princeton back to required testing, both effective for the 2027-28 cycle. That leaves six of the eight Ivies, Harvard, Yale (which ended its test-flexible policy in May 2026, closing the AP/IB alternative and moving to a straight SAT/ACT requirement), Dartmouth, Brown, Cornell, and Penn, requiring a score from applicants this cycle, a tier most IEC clients target directly, which means "should we submit a score" isn't really an open question anymore for those schools.

A short update: the Early Decision antitrust case just cleared a major hurdle

This is a different story from the testing consolidation above, and a different case from the 568 Presidents Group financial-aid settlement covered in a previous issue. On August 7, 2026, U.S. District Judge Angel Kelley in Massachusetts ruled that an antitrust lawsuit against 32 selective colleges over their use of Early Decision can proceed to discovery, according to Higher Ed Dive’s August 11 report and a same-day press release from Cohen Milstein, counsel for the plaintiffs. The case, D’Amico v. Consortium on Financing Higher Education, was brought not by a government agency but by four current and former students from Wesleyan, Vassar, and Washington University in St. Louis. The plaintiffs allege that the schools have an informal agreement not to recruit or admit students who break an Early Decision commitment, and argue that this coordination, not an independently enforceable contract between a student and a college is what gives ED its practical binding character. Judge Kelley found that claim plausible enough to survive the schools’ motion to dismiss, and in a separate ruling issued the same day, rejected a jurisdictional challenge from schools based outside Massachusetts, including Duke and Vanderbilt, keeping the full defendant list in the case. She did dismiss Common App, Scoir, and the Consortium on Financing Higher Education from the case, a result Common App’s own counsel, Steptoe, confirmed in an August 10 release, finding the plaintiffs hadn’t adequately alleged those organizations joined or facilitated the alleged conspiracy. The case now moves to discovery. Nothing about how ED works has changed yet, but the legal question of whether its binding character rests on anything more than an informal agreement among competitors is now formally in play, and it’s worth watching before you next describe ED as simply binding in a family meeting.

Knowing that ACT’s ownership just changed hands in the middle of its least stable technical stretch, at the exact moment testing became mandatory again at a narrow but high-stakes tier of schools, is one thing. Knowing what to actually tell a family building a testing plan this fall is what the paid section below covers.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What the Higher Education Act’s professional judgment provisions actually authorize an aid administrator to change, and why the section number many IECs have memorized for dependency overrides stopped being correct in 2024

• Why a financial aid administrator’s decision cannot be appealed to the Department of Education once it’s made, and how that authority is actually structured differently from school to school

• What a 2023 federal audit of one university’s professional judgment practice actually changed, and how narrow that shift really was across the schools that responded to it

For paid subscribers:

• What four published professional judgment policies actually accept and rule out, including the federal guidance’s own list of income adjustments the Department calls unreasonable, and one school’s worked examples of exactly who a documented loss actually helps

• The specific, sometimes conflicting documentation two of those schools each require for job loss, divorce, one-time income, and unreimbursed medical expenses, pulled from their own current appeal forms

• A structural framework for the appeal letter itself, built from the difference between the requests that get approved and the ones that get filed and denied without dispute

Upgrade to Paid

What professional judgment actually authorizes, and the citation most guidance gets wrong

“Appeal” is the word families use. The Higher Education Act uses a different one: professional judgment. Section 479A of the HEA, codified at 20 U.S.C. § 1087tt, gives a school’s financial aid administrator the authority to adjust, case by case and with adequate documentation, the cost of attendance, the values used to calculate the Student Aid Index, or the values used to calculate Pell Grant eligibility, for a student with special circumstances. It isn’t a general license to award whatever amount a family says it needs. It’s authority to adjust specific inputs, and only when documented circumstances support the adjustment.

That covers most of what families mean when they say they want to appeal an offer: a job loss, an unreimbursed medical bill, a divorce. A second, narrower authority covers dependency status. If a student’s situation makes it inappropriate to require parental information on the FAFSA, an FAA can override that status under what the statute calls unusual circumstances. For two decades, guidance referred to this authority as the seventh category, tied to Section 480(d)(7) of the HEA. A 2003 Department of Education letter to financial aid offices used that exact label. The FAFSA Simplification Act restructured Section 480(d) starting with the 2024-25 award year, and the Federal Student Aid Handbook now cites the authority as Section 479A(c) and Section 480(d)(9). The older number still shows up on well trafficked financial aid reference pages more than two years later, a small but telling sign of how little outside attention this corner of the law gets.

The statute is specific about what counts as a special circumstance, and specific in a way that should shape how a request gets framed. It names recent unemployment, a family member who qualifies as a dislocated worker, a change in housing status resulting in homelessness, an unusual amount of claimed losses against income, a severe disability, tuition at an elementary or secondary school, medical, dental, or nursing home expenses not covered by insurance, child or dependent care costs beyond the standard allowance, and an additional family member newly enrolled at least half time in a degree program. It closes with a catchall for other changes in income, assets, or family size. Two structural rules bound all of it. The circumstance has to differentiate one student from a group, not describe a condition shared across a whole class of applicants, and a school is barred from maintaining a blanket policy of denying every adjustment request of a given type. Since the FAFSA Simplification Act took effect, schools have also been required to publicly disclose, somewhere a family can find it, that this option exists at all.

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

Why the decision sits with the financial aid office, not with anyone above it

The feature of professional judgment that surprises most families, and a fair number of the IECs advising them, is what happens after a request is denied. In an admissions appeal, a family can sometimes escalate: write to the dean, ask for a second reader, push the department to reconsider. Professional judgment has no federal equivalent. The Federal Student Aid Handbook is explicit that once an FAA rules on an adjustment, that ruling is final and cannot be appealed to the Department. There’s no federal ombudsman, no secretary-level review, no path upward once the school’s own process has run.

How that process is structured inside a given school varies. The statute assigns the authority to the financial aid administrator, and at some schools that plays out as a single advisor’s call. At others it runs through a committee: Virginia Tech states plainly that its appeals are reviewed by a committee of its financial aid staff, not a lone gatekeeper. Either way, the discretion only moves in response to documentation that fits a recognized category. A well argued letter that doesn’t map to one of those categories isn’t weighed on a scale that tips with better writing. It’s measured against a file the office has to be able to defend if the school is ever audited, and that detail has mattered more over the past few years than it used to.

What a 2023 audit actually changed, and how narrow that change was

The audit that raised the stakes wasn’t about a small school. In August 2023, the Department of Education’s Office of Inspector General published a review of the University of Southern California’s use of professional judgment. Investigators found that USC had not applied professional judgment in accordance with Section 479A for 75 of the 108 students in its sample who received an adjustment other than a dependency override, across the 2019-20 and 2020-21 award years. USC disputed the findings and argued in its response that the Inspector General had applied a standard stricter than the statute requires. The Department did not walk the findings back.

The sector's reaction, when NASFAA measured it two months later, was narrower than the audit's headlines might suggest. NASFAA surveyed its membership in September 2023 and published results that October, receiving responses from 370 respondents at unique institutions out of a pool representing 2,717 institutions, a 14% response rate. 36% percent said they weren't familiar with the USC findings at all, and 31% said their institution was waiting for a final determination before deciding whether to act. Only 9%, roughly 27 institutions, said they were already changing or considering changing their professional judgment policies because of the audit. Within that small group, the shift was real: 67% reported limiting which circumstances they’d accept for a dependency override, and 26% said they expected to deny professional judgment requests going forward that they would have approved under their prior policies.

That’s a narrow slice of the sector, not a wholesale shift. But it’s the clearest documented evidence available that the audit changed behavior somewhere, and everywhere it did, the change ran the same direction: more documentation, less benefit of the doubt for a request that sits close to the line. Which specific circumstances still clear that bar, and which get filed and denied without any real dispute possible, differs by school in ways that are actually visible if you know where to look.

The mechanics above are the part every IEC should understand. But knowing that professional judgment exists is different from knowing how to use it.

For paid subscribers, we go one level deeper: the actual policies schools publish, the circumstances they accept and reject, the documentation they require, and the structure of an appeal that gives an aid administrator something they can actually act on. The goal isn’t to write a more persuasive letter. It’s to know exactly what belongs in the file, and what should never make it into the request in the first place.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Read “A National Call to Action to University Presidents and Governing Boards.”

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What the Education Department’s August 3 letter actually asks every college to publish by December 31, and why admissions criteria is the first item on the list

• Why this letter is a deliberately different animal from the Compact that seven of nine targeted schools rejected last fall, and the one school that answered it before it was even sent

• How the sector reacted in the first three days, and what the range of those reactions already tells you

For paid subscribers:

• A four-factor framework for scoring exactly how much federal attention a specific school on a client’s list is already carrying, using nothing but public record

• What to actually say when a family asks whether any of this affects their kid’s chances

• Why a binding Early Decision commitment can lock a family in before a school answers this letter, and why that risk doesn’t apply evenly across every school carrying it

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What the letter asks for, and why admissions comes first

On August 3, Secretary of Education Linda McMahon sent every college and university president in the country a letter titled “A National Call to Action to University Presidents and Governing Boards.” It asks each institution to publish, on its own website, a public statement addressing seven areas by the end of 2026: transparency and merit in admissions, free speech and open inquiry, intellectual pluralism, affordability and student outcomes, academic rigor in the age of AI, research integrity, and what the letter calls American interests. The letter itself runs three pages.

Admissions comes first on the list. The letter’s own text asks how a school will “ensure that admissions criteria are transparent and wholly described to prospective applicants” and how it will “ensure admissions decisions are based on merit, achievement, and your university’s educational purpose.” It does not name the Supreme Court by case name, but legal analysts reading the letter, including the law firm McGuireWoods, have connected that merit-based language directly to the Court’s 2023 ruling in Students for Fair Admissions v. Harvard, which ended the use of race as a factor in admissions. That reading fits the broader pattern of the last three years. What’s new is a federal request that every school say, in writing, on its own website, how its process holds up, rather than leaving that judgment to an admissions page written for prospective families.

The request is voluntary. Politico reported that the Department “did not immediately respond to questions about what happens if the schools do or do not participate,” and Foley Hoag’s client alert notes plainly that the letter carries no new legal mandates, no stated penalties for noncompliance, and no funding tied to participation. Legal analysts advising universities still read it cautiously. McGuireWoods, in its own alert, described the letter’s framing as suggesting that institutions which opt out may face more scrutiny, not less, and Inside Higher Ed reported a White House official telling Bloomberg that schools which do announce reforms could be more likely to be included in future conversations about federal research funding. Nothing in the letter’s own text makes either of those things explicit. For a family building a college list this month, the honest version is that a school’s response, or its silence, is being watched, even without a formal consequence written down anywhere.

Why this isn’t the compact almost every school rejected

This is not the first time in the last year Washington has asked universities to put commitments in writing, and the history matters for reading this one correctly.

Last October, the White House sent nine universities, MIT, Penn, the University of Virginia, the University of Arizona, UT Austin, USC, Vanderbilt, Dartmouth, and Brown, a nine-page document called the Compact for Academic Excellence in Higher Education. Signing it meant preferential access to federal grants. It also meant agreeing to specific changes, including a five-year tuition freeze, a cap on international undergraduate enrollment at 15%, and the reinstatement of standardized testing. Universities had until October 20 to respond and until November 21 to decide. Seven of the nine rejected it outright, University of Arizona among them, with its president writing that parts of the compact deserved consideration but that the university had not agreed to its terms. No university ever signed it. Two schools, Vanderbilt and UT Austin, never rejected it either. Vanderbilt’s chancellor told the campus the university had been asked to provide feedback rather than an up-or-down answer, and continued sending comments to the Department for months afterward. By January 2026, Secretary McMahon confirmed in a separate interview that the original document had been a preliminary draft, not a final one, and that a second version was being developed. As of this writing, no revised compact has been made public.

August 3’s letter shares a subject with the Compact and almost nothing else. It carries no signature line and no funding attached to compliance. The Department consulted with groups including the Association of American Universities before sending it, itself a break from how the Compact arrived with no warning. Ted Mitchell, president of the American Council on Education, called the Compact a “Faustian bargain” last fall. Asked about this letter, he described it instead as “an invitation to a conversation.” Whether that softer framing produces different results is the open question, and it’s worth noting that this is exactly the language Vanderbilt’s own chancellor has been using to describe his university’s relationship with the Department since October.

That posture may not be a coincidence. Five days before the letter went out, Vanderbilt’s Board of Trust published “The University and Its Purpose: A Declaration of First Principles,” a document it had quietly adopted a month earlier, on June 29. Vanderbilt and Washington University in St. Louis had also jointly commissioned a separate committee, led by philosopher Paul Boghossian, that reported in June on what it called the politicization of several humanities disciplines. McMahon’s own letter references that committee’s work directly, by name. Whatever coordination existed between Vanderbilt and the Department before August 3, Vanderbilt’s declaration already covers admissions standards, free expression, foreign influence, and academic rigor, largely the same ground the Department’s letter now asks every other school to cover from scratch. The administration appears to be applying the same open-ended posture to every school at once that it has already used for ten months with Vanderbilt specifically. The difference is that Vanderbilt had already put something on paper before being asked directly.

The first three days: measured, not alarmed

The reaction so far tracks the letter’s own tone. Ross Mugler, president of the Association of Governing Boards of Universities and Colleges, described the seven priorities as goals the sector already shares, while noting that institutions need room to pursue them in their own way. DeRionne Pollard, head of the American Association of Community Colleges, made a narrower point worth passing on to any family considering a community college transfer path: several of the seven areas, selective admissions and research security among them, simply don’t describe how most two-year institutions operate, and her members aren’t treating the letter as something they need to respond to.

No university has published anything formally framed as a response to this specific letter. Vanderbilt is the partial exception, though not because it moved quickly this week. Its board did that work in June for its own reasons, and the timing happened to land five days ahead of the Department’s letter. Reaction to Vanderbilt’s declaration has split along familiar lines. Heterodox Academy and a Stanford free-speech alumni group praised it as principled leadership, while an Inside Higher Ed opinion columnist called it a threat to faculty free expression. At the University of Arizona, one of the original nine Compact recipients, the president has said he’ll consult shared governance bodies before deciding on next steps, a pattern likely to repeat at most other schools weighing how seriously to treat a request with no deadline penalty attached to it.

None of that tells you whether a specific school on a specific client’s list is one of the ones under real pressure right now, or how to answer a family who has already seen a headline about this and wants to know what it means for their kid. That’s what the rest of this issue does.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What actually carried over from the 2025-26 cycle into this one, and what quietly reset the moment the clock struck August 1, including the renamed Activities and Experiences section and the newly optional title field sitting inside it

• Why application volume kept accelerating through the back half of last cycle specifically, and what that curve means for how firm any list is in September

For paid subscribers:

• The list size framework built for a cycle where adding a school no longer costs every student the same thing, broken out by selectivity tier and by which additions are now close to free

• How to coach the optional title field by student profile, including where the old advice about manufacturing a title was actually protecting a student and where it was never doing anything at all

• The tracking system for running a caseload you have no counselor portal visibility into, built around a Common App feature most IECs invite too late or not at all

• The full pre-Labor-Day checklist covering rollover verification, recommender re-invitations, supplement re-audits, and testing policy re-checks, in the order that actually matters

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What carried over from June, and what reset the moment the clock struck August 1

A rising senior you started working with in June logs into her Common App account this week and sends you a screenshot, confused. The section she filled in over the summer, the one she still calls Activities, is now labeled Activities and Experiences. The two teachers she invited to write recommendations in July are gone from her account entirely. The supplement she drafted for a target school over the spring is gone too, along with the prompt it was answering. None of this is a glitch. All of it is by design, and if you are fielding a version of this message from more than one client this week, the design is worth understanding in full.

Common App’s first-year application for the 2025-26 cycle closed on July 28. The platform went dark for three days and reopened August 1 for 2026-27, a reset it runs every year on the same schedule. What makes this year’s reset worth a dedicated walk-through is less the reset itself than the specific list of what does and does not survive it, because that list determines how much rework your August sessions actually require.

• What survives: a student’s username and password, her college list, her answers across the core Profile, Family, Education, Testing, and Activities sections, and any advisor invitations she has already sent. Advisors are a distinct category from recommenders. An advisor can view a student’s application without submitting anything to it, and Common App has confirmed that advisor invitations, unlike recommender invitations, persist across the rollover rather than resetting. More on why that distinction matters below.

• What resets to zero: every recommender invitation, meaning any teacher or counselor a student invited before August 1 has to be reinvited from scratch. The FERPA release authorization resets and has to be resigned. Every college specific supplemental question resets, along with a student’s response to the Common App fee waiver question. Any question whose wording changed year over year resets even within the core application. A student who spent July polishing a supplement for a school that has since revised its prompt is not polishing a draft anymore. She is starting over on a prompt that no longer exists.

• Layered on top of the reset is the actual content of the 2026-27 application, and the headline change is the renaming of the Activities section to Activities and Experiences. The character limits underneath it are unchanged, 50 for the position or leadership title, 100 for the organization name, 150 for the description, still across up to 10 entries. What changed is that the position and leadership field is now optional rather than required. Common App also relabeled several activity categories to better match how students actually spend their time outside class. Work (paid) is now Employment or work (paid). Junior ROTC is now ROTC and Military. A new dedicated field captures post-graduation activities and plans, giving students a cleaner place to note a gap year or a confirmed commitment after high school without forcing it into the 10-entry list. And the U.S. resident status question, which for years generated confusion among American-born applicants who misread the label, now carries clearer wording and a help prompt.

• None of this touches the personal statement. Common App confirmed on February 27 that the seven essay prompts carry over unchanged into 2026-27, so any draft a student built over the summer against this year’s prompts is still answering a live question. The Additional Information field, cut from 650 words to 300 in the prior cycle, stays at 300 words this year. If your intake process leans on last year’s guidance about that section, the guidance still holds.

The platform itself kept growing. Common App’s own launch announcement puts total membership at over 1,200 institutions for 2026-27, and the accompanying language points to continued investment in what the organization calls access initiatives for students from low and middle income backgrounds. The clearest expression of that investment is the Direct Admissions program, which launched at full scale three cycles ago with 70 participating colleges and grew past 200 for 2025-26. Common App’s own data on the program shows a majority of participating colleges returning for another cycle, and some now waive essays or recommendations entirely for students who receive a direct offer. A handful of states, including Connecticut and Illinois, have folded their own statewide automatic admission programs directly into the platform. We will come back to what this means for how many applications a given student actually needs, because it changes the math more than most list-building conversations currently account for.

Why application volume kept accelerating through the back half of last cycle

The volume story from the 2025-26 cycle is not just that applications kept climbing. It is that the climb sped up as the season went on, and the shape of that acceleration matters more than the final number.

Through November 1, 2025, Common App recorded 962,284 first-year applicants who had submitted 4,716,352 applications to 916 member institutions, an average of 4.90 applications per applicant, up from 4.68 at the same point the year before, roughly a 5% jump. A month later, through December 1, the applicant count had grown to 1,158,805 and total applications to 6,237,325, pushing the per-applicant average to 5.38, up from 5.15, a total application volume increase of 9% year over year. By March 1, 2026, the average had reached 6.59 applications per applicant, up from 6.41 at the same point the prior cycle, with total applications through that date crossing 9.4 million.

Lay those three checkpoints next to each other and the pattern is not a steady climb. It is an average that grew by roughly 0.48 applications per applicant between November and December, then by roughly 1.21 applications per applicant between December and March, close to two and a half times the rate. Most of a season’s growth in list size happens after the early deadlines have already passed, not before them. A list a family considers finished in October is, for a meaningful share of applicants, not the list they actually submit in February.

The growth was not distributed evenly across the applicant pool either. First-generation applicants grew 12% compared to the same point the year before, more than five times the 2% growth among continuing-generation applicants. Students eligible for a Common App fee waiver grew faster than students who were not. Underrepresented minority applicants grew at roughly twice the rate of the overall pool. This is a continuation of a pattern this newsletter has tracked before in the context of the activities disparity research, where lower-income and first-generation students have historically appeared with fewer entries and fewer formal titles on their applications, not because they did less, but because the format rewarded a kind of record-keeping some households have more practice producing than others. The fastest-growing segment of this year’s applicant pool is exactly the segment for whom an extra application, an extra supplement, an extra deadline has historically carried the highest relative cost in time and confidence.

International application volume moved in the opposite direction, down 9% overall compared to the same point last cycle, with steeper declines from Asia and Africa specifically and a 14% drop in applicants from India, a trend that connects to the visa and immigration coverage we’ve published this year and that likely continues showing up in your international families’ calculus this fall.

More member institutions, an expanding Direct Admissions program, and a growing number of schools that dropped their supplements this cycle have all lowered the marginal cost of adding one more school to a list. At the same time, the applicant pool most likely to benefit from that lower cost is also the pool historically most vulnerable to overextending into more applications than they can finish well. Lower friction is not the same thing as a better outcome. What actually constitutes the right number of applications for a given student this cycle is not a question the volume data answers by itself. It’s a decision, and the framework for making it deliberately, rather than by default, is where the paid section below picks up.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• Why the College Board killed Landscape, the socioeconomic data tool colleges used after the Supreme Court banned race-conscious admissions, and what its statement actually admitted

• How a federal data collection built specifically to catch race by proxy is now working through the courts, and which schools it currently reaches

• What Georgetown told the public about its own Pell eligibility flagging, and why that statement makes it the clearest test case anywhere for whether this kind of proxy survives scrutiny

For paid subscribers:

• A framework for checking which schools on a client’s list are likely running a version of what Georgetown disclosed, using the same public filings this newsletter has already taught you to read

• What changes for families now that the tool that used to supply context automatically is gone, and what a student needs to put in the application themselves instead

• The specific intake conversation for any client applying to Georgetown this cycle, covering the platform choice, the testing requirement, and what stays exactly the same

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

Don't take our word for it. Here's what the higher ed community thinks.

What Landscape did, and why it is gone

From 2016 until last year, the College Board offered admissions offices a tool called Landscape. An applicant’s home address and high school fed a dashboard showing median family income, the share of single parent households, and other neighborhood-level indicators, alongside a relative disadvantage score built from that data. It never touched race or ethnicity. Colleges used it to identify strong applicants from under-resourced schools, and some used it to shape financial aid offers once those students were admitted.

Last September, the College Board confirmed it was discontinuing Landscape. Its stated reason was brief: federal and state policy on how institutions use demographic and geographic information in admissions continues to evolve, so the tool would end. The statement repeated that Landscape had been built without any use of race or ethnicity data, a defense that mattered because the tool had drawn direct fire from Students for Fair Admissions, the group whose litigation ended race-conscious admissions at the Supreme Court in 2023. Edward Blum, the group’s founder, told the New York Times that any tool letting admissions offices weigh race by proxy carries legal and reputational risk. Richard Kahlenberg, who served as an expert witness for Students for Fair Admissions during that same litigation and has spent years advocating for class-based alternatives to race-conscious admissions, called the College Board’s decision to abandon Landscape unfortunate, since the tool was race-neutral and its use was legal.

The disagreement between those two men, both aligned against race-conscious admissions, over whether Landscape itself should survive is the story. It shows that removing race from a formula did not settle what colleges are still allowed to know about the disadvantage a student overcame to get in front of them. The tool’s own developers said it was clean. It got shut down anyway, in an environment where any correlate of race, however indirect, has become a liability an institution weighs against the benefit of using it.

The federal audit built to find what replaced it

Landscape did not disappear in isolation. In August 2025, President Trump issued a memorandum directing the Secretary of Education to expand federal data collection from colleges within 120 days, specifically to increase transparency into admissions. Secretary Linda McMahon directed the National Center for Education Statistics to act on that timeline, and by mid-December the Office of Management and Budget had approved a new component of the annual federal reporting system every degree-granting college already files: the Admissions and Consumer Transparency Supplement, known as ACTS.

ACTS asks four-year colleges with selective admissions to report seven years of applicant-level data, broken out by race and sex and further cross-tabulated by test score range, GPA range, family income band, Pell eligibility, and whether a student is the first in their family to attend college. The Federal Register notice announcing it was direct about the purpose: the data is meant to reveal whether an institution is using diversity efforts as a pretext for the kind of racial preference the Supreme Court barred in 2023, including preferences applied through indirect proxies rather than race itself.

Colleges had roughly three months to compile historical data most had never assembled in this form, and the rollout has been contested since the first deadline arrived. On March 11, seventeen states led by Massachusetts sued the Department of Education, arguing the survey was implemented so quickly that it violated federal rulemaking law. A federal judge in Massachusetts has since granted preliminary relief in stages: first for public colleges in those seventeen states, then, in a second ruling on April 24, for roughly 178 additional private institutions represented by six higher education associations, including the Association of American Universities. Georgetown is a member of that association. In both rulings the court found the Department of Education likely has the legal authority to collect this data, and that the survey was probably rolled out in an arbitrary and unlawful manner given how little time institutions were given to prepare. The court declined to block the requirement nationwide, so it currently still applies to more than two thousand four-year colleges that were not party to the lawsuit, while the litigation over the rushed rollout continues. The Department of Education has said the data is needed to confirm compliance with the 2023 ruling. Data from the first collection cycle is expected to become public in the coming months, and once it does, according to a July analysis in Inside Higher Ed, colleges that have not already prepared an explanation for their own admissions patterns should expect the numbers to reopen the debate over who gets in and why.

The one school that already told you the answer

Set against that backdrop, Georgetown’s own account of its process reads differently than it did when the university first published it. In March 2025, Georgetown announced the three-year Common App pilot alongside two other decisions worth reading closely now. It would keep requiring SAT or ACT scores from every applicant, citing new National Bureau of Economic Research analysis showing that test-optional policies disproportionately hurt applicants from under-resourced high schools. And it confirmed something it had actually begun a year earlier: since 2024, Georgetown has factored an applicant’s Pell eligibility into its comprehensive admissions review, specifically to grow the share of low-income students on campus.

The university was equally direct about the outcome. Its Class of 2028, the first admitted after the Supreme Court’s ruling, had the highest share of Pell-eligible students Georgetown had enrolled in more than a decade, at 15 percent. The university has since raised its financial aid budget to 285 million dollars for the coming year, its largest ever commitment, and has said explicitly that the increase is meant to help keep growing Pell enrollment.

Few peer institutions have described their admissions process this explicitly in a public announcement. Most institutions describe class-based diversity work in general terms, comprehensive review, context, opportunity. Georgetown named the mechanism. It said it looks at Pell eligibility, said when it started, and said what it is trying to accomplish by doing it. That level of specificity is exactly what a Common Data Set audit or an ACTS submission is designed to surface for every other school whether or not they choose to say it publicly. Georgetown already has. It becomes the reference case against which every other selective college’s practice will now get measured, at the same moment its applicant pool is about to expand through a platform used by more than a million students a year, a large share of them first-generation and lower income, according to Common App’s own figures.

The federal audit exists to identify admissions practices that may rely on indirect racial proxies. Georgetown has publicly described one such socioeconomic factor it considers, making it a natural test case for how these practices are viewed. Whether Pell flagging survives that scrutiny as a legal, workable practice, or becomes the next Landscape, is a question this newsletter expects to have a real answer to well before this three-year pilot ends.

The paid section below walks through how to check whether a specific school on a client’s list is likely running something similar to what Georgetown disclosed, what changes for a student’s own application now that a tool used to supply context automatically is gone, and the specific intake conversation to have with any family applying to Georgetown this cycle.

What you cannot yet tell is which other schools on a client's list are already doing a quieter version of what Georgetown put in writing, and there is a way to check that does not require waiting for a press release. The paid section below shows exactly where to look, what now has to be said explicitly in a counselor letter that used to be implied, and the specific intake script for every Georgetown applicant on a caseload this fall.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Think of this as a standing reference rather than a one-time read. Bookmark it, search it when a client question echoes something we’ve covered before, and send it to a colleague who’s newer to the field and could use a fast way to get oriented. We’ll publish an updated version when we hit 100 issues.

1. Federal Policy, Financial Aid & Institutional Finance

How 477,000 missing FAFSA completions during the botched 2024-25 rollout erased a generation of applicants, concentrated overwhelmingly among low-income, first-generation, and minority students. Covers what financial aid directors are seeing in 2026 yield data, the financial markers separating colleges likely to survive the next 18 months, and previews the Parent PLUS cap arriving July 1.

DOGE’s cancellation of $881 million in Institute of Education Sciences contracts, including the loss of NPSAS and other federal data infrastructure that financial aid and enrollment offices rely on for benchmarking. Includes a six-step response plan for institutional research offices and guidance for rebuilding benchmarks without federal data.

This piece breaks down the One Big Beautiful Bill Act’s Parent PLUS cap ($20,000/year, $65,000 lifetime) and the elimination of Grad PLUS loans, both effective July 1, 2026, plus the fight over which graduate programs count as “professional” for higher borrowing limits. Flags a hidden repayment trap for existing Parent PLUS borrowers and gives a pre-June 30 action checklist.

How the private-college tuition discount rate hit a record 56.3%, turning “merit scholarships” into an unsustainable prisoner’s dilemma that has already pushed several colleges to closure. Gives counselors a practical framework, including free tools like College Viability and IPEDS, for screening a school’s financial health before recommending it.

How core Department of Education functions and staff are being quietly reassigned even as the agency nominally stays open, and what that structural shift means for financial aid processing and oversight going forward.

Post 22 Follows directly from the Parent/Grad PLUS caps to examine the private lending market set to roughly double in volume: the gaps between federal and private loans, the credit-based approval divide, the “starting at 2.69%” variable-rate trap, and a full framework for evaluating any private loan offer before a family signs.

This piece explains the Department of Education’s new “Do No Harm” earnings test, the first accountability rule to apply one standard to every postsecondary program at every institution type, and which fields (cosmetology, early childhood education, graduate social work and mental health counseling) are failing at the highest rates.

This piece details the One Big Beautiful Bill Act’s tiered endowment tax, already costing schools like Harvard and Notre Dame tens to hundreds of millions of dollars a year, and connects it to the June antitrust case to question what “meets full need” claims are actually built on.

A practical follow-up once the Parent PLUS cap actually took effect July 1: the SAVE repayment deadline, who it affects, and how to rebuild a client’s college list under the new borrowing caps.

Why the CSS Profile, not the FAFSA, is the number that actually determines affordability at the roughly 200 mostly private colleges that require it, and why families targeting Early Decision at CSS schools need their real aid estimate the same month they must commit.

Upgrade to Paid

2. Enrollment Trends, Institutional Stability & Alternative Pathways

Four straight years of community college enrollment growth against private four-year decline, driven by families doing straightforward math on total cost. Models three scenarios through 2030 for which institution types survive the shift.

Pulls together May 2026 graduate outcomes data, the federal funding fight hitting research universities, and a cascade of nonprofit college closures (8 so far in 2026, with 440 more projected at risk) into a single framework for evaluating whether a specific degree at a specific school still justifies its cost.

This piece (co-authored with , Katz College Planning) covers “degree hacking”: CLEP exams, competency-based programs, and 90-credit bachelor’s degrees letting students compress four years into a fraction of the time. Gives IECs a framework for telling credible accelerated pathways from risky ones.

3. AI in Admissions & Counseling

This piece opens our AI coverage with the reality that a 408:1 national student-to-counselor ratio has pushed families toward ChatGPT as a default advisor, producing both genuine access wins and a wave of essays that read, in one admissions officer’s words, like “a robot trying to remember what being a teenager felt like.”

Introduces Generative Engine Optimization: how families are increasingly finding IECs through ChatGPT and Perplexity rather than Google, and the specific content and platform tactics that determine whether a practice gets cited by AI tools or stays invisible.

This piece shifts the focus from students to practices: AI disclosure questions, the privacy exposure most IECs don’t realize they have, application clustering risk, and a full data-hygiene and tool-selection protocol for counselors using AI in their own workflow.

One of our most-read pieces. Uses Gallup/Lumina and Oxford Internet Institute data to separate which majors, institution types, and career pathways are genuinely losing ground to AI from which just feel that way, with a field-by-field advising framework for families this summer.

Handshake data shows 85% of the Class of 2026 used AI tools in college despite only 28% receiving formal instruction. Covers Purdue’s first-in-the-nation AI competency requirement and what families should look for in real institutional AI programming versus marketing talk.

New research, including a Stanford study showing AI-detector false-positive rates jump from 5% to 61% on non-native English writing, on what AI use is doing to admissions essays across an entire applicant pool. Lays out a coaching workflow for protecting a student’s actual voice once a prompt is chosen.

Traces the widely repeated “half of all colleges use AI in admissions” statistic back to a marketing survey a sitting vice provost has directly called false, and instead documents the small number of institutions that have actually disclosed how AI touches their review process.

Upgrade to Paid

4. Running an IEC Practice

A conversation with , IECA’s CEO for thirty years, on how thirty years of professionalizing independent educational consulting revealed, rather than closed, a two-tier gap between families with access to expert guidance and those without. Includes five business models for extending access without breaking practice economics.

Runs the actual numbers on why an IEC based in a lower cost-of-living market serving premium regions like Westchester or the Tri-State corridor remotely can out-earn, in real purchasing power, a counselor based in and serving that expensive market directly.

A caseload-triage framework for sorting clients by actual time demand before an IEC gets underwater in October, so practices can price and staff realistically instead of discovering the imbalance mid-cycle.

Makes the case that the eight weeks between May 1 and late June are the highest-value client acquisition period on the IEC calendar, and that most practices let the window pass without a deliberate outreach plan.

Maps eight forces reshaping IEC practice for 2026-27, from a testing reversal at nine Ivy Plus schools to Common App architecture changes to AI-detection liability to Hampshire College’s closure, and translates each into specific pre-fall action items.

Introduces a downloadable six-dimension student assessment framework for the first client meeting of the cycle, covering why May and June are the last moment every advising decision is still fully correctable.

We use Common App’s record 9.4 million applications and 6.59-apps-per-student average to argue that memory-based practice management breaks under real caseload volume. Includes a self-audit and infrastructure framework for fixing it before the fall cycle compresses.

5. Admissions Strategy & The Application

We argue that Ivy Day headlines cover a sliver of the roughly 19.6 million students actually navigating the 31 days between decisions and May 1, and walks through the enrollment cliff, the merit aid war, and the financial aid comparison problem shaping decisions everywhere outside the top eight schools.

A snapshot of the financial pressures, antitrust settlement landscape, testing shifts, and deposit-season dynamics at selective institutions that counselors need to translate for families in real time.

Draws on IECA data showing independent consulting has grown more than fivefold since 2010 to argue that students form their real college preferences in social, informal channels, Reddit threads, WhatsApp groups, counselor referrals, long before they register in any institutional CRM.

Compares how Boomer, Gen X, Millennial, Gen Z, and Gen Alpha parents and students each experience the admissions process differently, and argues counselors who read which generational lens a family is operating from give measurably better advice.

Uses one family’s situation, a parent 44 payments from Public Service Loan Forgiveness, to show how a new Parent PLUS loan taken out after July 1 can retroactively strip repayment protections from a family’s entire existing balance. Gives exact language for the conversations that need to happen before enrollment and loan-disbursement windows close.

52% of Common App applicants submitted test scores in the 2025-26 cycle, the first submitting majority since before the pandemic, even though most schools remain officially test-optional. Covers the peer-reviewed research on who gets hurt when “optional” is misread as “doesn’t matter.”

Schools now fill 40-60% of their incoming class through binding early rounds. Covers the financial analysis that should precede any ED recommendation and a step-by-step readiness framework for families who are financially capable but not fully certain.

We make the case that the summer before senior year is the application, not preparation for it, with a week-by-week protocol from June 1 through August 15 covering recommender conversations, a 40-minute “Why Us” research method, and an activity-list rewrite session.

We use the NAEP, ACT, and NCES data to document the widening gap between record-high high school GPAs and record-low measured academic proficiency, and how admissions offices are already adjusting for grade inflation.

A full breakdown of demonstrated interest: a 3-minute Common Data Set lookup to check whether it matters at a given school, how CRM platforms like Slate actually track behavior, WashU’s new formal DI factor, and a summer outreach protocol running June 1 through August 15.

Addresses advising a family whose older child got into college under rules that have since changed. Argues that leading with a policy explanation is the wrong opening move and offers a reusable “what’s changed since” framework instead.

Explains why most students fill their ten Common App activity slots in the wrong order, what the new Responsibilities and Circumstances checklist changed, and gives a slot-by-slot sequencing and 150-character compression framework.

Uses Common App’s own prompt-selection statistics to show a concentration problem, most students default to the easiest prompt, and gives a prompt-by-prompt diagnostic plus three questions that surface a student’s strongest material before a prompt is chosen.

Draws on court records from the Harvard and UNC admissions lawsuits plus Jeffrey Selingo’s year embedded in three admissions offices to reconstruct how a file actually gets read, and what that means for the activity list, essay, and additional information box.

Six selective universities eliminated or reduced supplemental essay requirements for the 2026-27 cycle. Examines the stated “reduce stress” rationale against what’s actually driving the change and what it means for essay strategy.

6. Equity, Legal & Access Issues

One in five Gen Z students now searches for scholarships on TikTok weekly, more than consult a school counselor, and a third encounter content they believe is misleading: fake awards, data-harvesting “matching services,” and hidden processing fees.

Why 71% of admissions counselors leave their jobs within three years, how Georgetown joining the Common App is projected to spike regional rep workload 30-50%, and what durable access channels IECs should build instead of relying on any single relationship.

Covers the federal financial-aid antitrust case in which twelve of seventeen named universities have already paid roughly $320 million to settle claims they factored family wealth into admissions while claiming an exemption that required them not to. The five holdouts, Cornell, Georgetown, MIT, Notre Dame, and Penn, are fighting to delay a November trial.

Five states have now banned legacy admissions, two of which just hit their first compliance deadline. Clarifies exactly which schools on a typical client list are actually covered versus assumed to be.

Unpacks the practical difference between “test-blind” (UC’s actual policy) and “test-optional” as the UC Board of Regents opens a faculty-led policy review, and what that distinction should change about how families approach testing for UC applications specifically.

Upgrade to Paid

7. Special Populations & Global Systems

IEPs and 504 plans do not carry over to college. Walks through what actually replaces them (self-disclosure to a disability services office under the ADA rather than IDEA) and the summer documentation work that determines whether a family discovers this gap in September or in April.

A new DHS rule ends open-ended F-1 student status in favor of a fixed four-year admission period. Only 44% of first-time bachelor’s students finish within four years, and the piece audits which program types (architecture, engineering co-ops, combined degrees) put even domestic-track international students at risk of running past the new limit.

We look at India’s NEET protests, Nigeria’s JAMB collapse, and South Korea’s Suneung controversy as three case studies in the structural fragility of admissions systems built around a single high-stakes exam.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

That’s the full library so far. If there’s a topic you’ve been wanting us to dig into, a specific policy question, a corner of the admissions process, something happening with one of your clients that feels bigger than just their case, reply and let us know. Some of our best issues started as a reader’s question.

Leave a comment

Higher Ed Insights publishes weekly for IECs, families, and higher education professionals. Tips, story ideas, and feedback can be sent to the editorial address on our About page.

Free subscribers receive the problem framing. Paid subscribers receive the specific playbooks. If you are reading on a free subscription and found this piece useful, the market-by-market entry strategies and acquisition playbooks above are what paid looks like.

Share

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What actually changed in the AI-detection landscape since our essay-voice issue, and the one tool now doing something none of its predecessors could

• The recommendation letter blind spot: what AI can ethically touch in the brag-sheet-to-letter pipeline, and the exact point past which it becomes a different kind of problem

• A quick, practical note on AI notetakers, and why the honest answer here is simpler than the internet’s current AI-panic content would have you believe

For paid subscribers:

• The list-building workflow that produces a genuinely differentiated list instead of the same twelve schools every AI tool recommends

• The exact brag-sheet prompt to use with a student, and the line it must never cross into letter-drafting territory

• How to talk to a family about their own AI use, since most of them are already experimenting with it whether you bring it up or not

• A full prompt library: ready-to-use prompts for your own workflow, and a separate set built to hand directly to students and parents

• Our take: why marketing “we use AI too” may be undermining the actual case for hiring you

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

Don't take our word for it. Here's what the higher ed community thinks.

What actually changed since we last covered AI detection

When we wrote about essay voice and AI detection, the honest conclusion was that detection tools were unreliable enough that no family should build a strategy around one specific tool catching or missing anything. That conclusion still holds. What’s changed is that the market has stopped treating all detectors as roughly equivalent.

A University of Chicago Booth School of Business working paper, by researchers Jabarian and Imas, benchmarked the major AI detectors, including Turnitin, GPTZero, Originality.ai, and Pangram, against both raw AI-generated text and text run through “humanizer” tools designed specifically to evade detection. Pangram was the only detector that held a false positive rate near zero even after humanizer processing degraded every other tool’s accuracy. That result is doing real work in the market: Pangram has moved from a niche entrant to the tool that admissions-adjacent detection services and several institutions now treat as the reference standard, while Turnitin’s AI-detection product, disabled outright by Vanderbilt in 2023 over accuracy concerns, continues to lose institutional trust into 2026.

Worth being precise about, since this is exactly the kind of nuance that gets flattened in most AI coverage: GPTZero disputes this paper’s ranking, arguing the researchers queried the wrong field of its API and that on a corrected re-run it matches or beats Pangram. Read that dispute closely and it’s narrower than it sounds. GPTZero is contesting its own recall score, its ability to catch AI text at all, not Pangram’s false positive rate, the number that actually matters for a family worried about a human-written essay getting wrongly flagged. The two detectors may be closer on catching AI text than the headline numbers suggest. On wrongly accusing an honest student, the independent finding still stands unchallenged.

The detail worth carrying into a client conversation is not “Pangram is better.” It’s that Pangram’s output is more granular than a simple flag. It can distinguish AI-written from AI-edited text at something closer to the sentence level, which matters directly for the coaching guidance in our essay-voice issue: a student who used AI to reorganize a paragraph they wrote themselves is now more distinguishable, in principle, from a student who generated the paragraph outright, than detection technology allowed for even a year ago. That is a reason for tighter coaching discipline, not looser. The safest version of AI-assisted editing keeps the tool at the outer edge of a draft, structure, pacing, word choice at the margins, and never inside a sentence the student didn’t write first.

The recommendation letter blind spot

Every issue in this series has covered AI’s effect on work a student produces directly: essays, activity lists, brag sheets. None of it has addressed the piece of the file an IEC is closest to without ever touching: the recommendation letter.

You don’t write recommendation letters. That’s the point, and it’s exactly why this deserves scrutiny rather than a pass. IECs routinely help students prepare the brag sheets and resumes that get handed to teachers and counselors, the raw material a recommender draws from to write the letter. AI has made that preparation dramatically faster. It has also created a specific failure mode most IECs advising on this process haven’t been warned about.

Jeffrey Neill, director of college counseling at Graded: The American School of São Paulo, described the honest math of the job to EdWeek: a recommendation letter used to take him about three hours, and at least ninety minutes of that was spent collecting information from scattered sources rather than writing. That is the legitimate use case: AI aggregating a student’s transcript, brag sheet, and a counselor’s own notes into an organized summary the counselor then writes from. The AI never touches the letter. It only organizes material a human was always going to use, which is exactly the aggregation the American School Counselor Association’s own guidance already expects counselors to do before writing, gathering ample information about a student first, regardless of what tool does the gathering.

The line gets crossed when the aggregation quietly becomes drafting, when a brag sheet arrives so fully written, in such polished paragraph form, that a busy recommender is functionally being handed a draft letter to lightly edit and sign. A 2024 study on AI use in the college process found roughly one in three students and teachers had already self-reported using generative AI to help with essays or recommendation letters, and a co-author of the study has since said publicly she expects that number has only grown. Admissions readers who review hundreds of letters a cycle have documented pattern recognition for exactly what that produces: language that sounds plausible about a student without being specific to them, the same handful of adjectives, “resilient”, “curious”, and “hardworking” recurring because those are the words AI reaches for when it’s filling gaps rather than recalling a real memory.

This was already a known problem with brag sheets before AI entered the picture: college advisers have long warned that a generic, checklist-style brag sheet produces recommendation letters that read like every other recommendation letter, because every recommender working from the same template reaches for the same language. AI does not introduce that failure mode. It makes it much faster to produce at scale, and much harder for a recommender to notice they’ve fallen into it.

Here is the part of this that should actually change how you talk to a student about their brag sheet, not just how a recommender uses it. A student has no control over whether their counselor or teacher ran the finished letter through an AI tool, no visibility into it, and no way to defend against a flag if the letter reads like AI output. A student can write a flawless, entirely human essay and still end up with a file that raises questions, because of a choice a recommender made in a different part of the application the student never saw. It’s a real reason to make sure the brag sheet you help a student build gives a recommender enough specific, human detail that they don’t feel pressure to lean on AI to fill in what’s missing.

The standard worth holding, and worth explaining to a student before they hand a brag sheet to a teacher: AI can help a student remember and organize what actually happened, a specific class discussion, a specific project, a specific moment a teacher would recognize, but it should never generate the descriptive language a recommender is expected to supply from their own memory and judgment. This isn’t a new rule invented for AI. It’s the same confidentiality and honesty standard IECA’s own Principles of Good Practice already hold this profession to, applied to a tool that didn’t exist when those principles were written.

A quick, practical note on AI notetakers

Otter, Fireflies, Fathom, and the built-in transcription tools in Zoom and Teams have become close to standard in professional meetings generally, and plenty of IECs now use one to avoid typing notes by hand during a family call. To be direct about it: this is a sensible, low-risk thing to do, and the case for treating it as a serious legal exposure is thinner than a lot of current AI commentary suggests. A family who hired you is not adversarial to you, and the realistic path from “my consultant used a notetaker” to any actual dispute is close to nonexistent for the overwhelming majority of practices.

The one habit worth building anyway, because it costs nothing and removes any ambiguity: say so upfront. When a new family starts working with you, mention during intake or in your engagement letter that you may use an AI note-taking tool on calls so you can stay focused on the conversation instead of your notepad, and that they’re welcome to ask you to turn it off at any time. One sentence, said once at the start of the relationship, does essentially all the protective work a much longer conversation would otherwise be trying to do.

Everything above is diagnosis: what changed, where the blind spot is, what the actual risk looks like. Below the paywall is where this issue turns into something you can use immediately: the list-building workflow that keeps AI from handing every family the same twelve schools, the exact brag-sheet prompt and the line it can’t cross, how to actually raise AI with a family, and a full prompt library built for your practice and ready to hand straight to students and parents.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Unrelated to the piece below, but worth a watch:

This video reports some of the important topics we have covered previously, including: Grad PLUS new limits, OBBB, access to grad programs, etc.

“Grad PLUS loans were created in 2006, allowing graduate students to borrow up to the full cost of attendance. Total borrowing rose from about $2 billion in the program’s first year to over $15 billion in 2024–25, but beginning July 1, 2026, new borrowers will face annual loan caps, leaving a funding gap for students and universities alike. Supporters say the limits will curb overborrowing and tuition growth; critics warn they could push students toward private loans and reduce access to lower-paying fields.”

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What the international enrollment collapse actually looks like this year, told through the institutional numbers behind the headlines

• Why a story about foreign enrollment turns into a story about your clients’ financial aid, once you follow where the money actually comes from

• The specific programs, positions, and tuition increases that have already disappeared because of this, at a growing list of named institutions

For paid subscribers:

• How to check any school on a client’s list for this exact financial exposure before this fall’s applications go in

• What that exposure predicts about which majors and programs are actually safe to build a list around, and which ones deserve a second look

• The specific questions to ask a financial aid or admissions office, and what it means when the answer is silence

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

Keller was starting his second year as president of the University of North Texas last fall when the number came in. Full-tuition-paying international students, especially graduate students, who Keller has said bring $20,000 to $25,000 each to the university’s bottom line, underwrite institutional services, staffing, and domestic financial aid. When 2,800 of them didn’t show up, UNT lost the exact revenue cushion that had previously offset state budget pressures.

“I picked a hell of a time to become a college president,” Keller said.

UNT eliminated 71 academic programs. Its budget office projects another $47 million in lost revenue next year, though Keller expects the actual hit to land closer to $25 million given the cuts already underway. None of this made national news the way a campus protest or a visa raid does. It is, instead, the kind of story that shows up quietly in a family’s acceptance letter, in a program that no longer exists in the form it did when a student applied to it, or in a financial aid package that is smaller than the one a sibling got two years ago at the same school.

What the international enrollment collapse actually looks like this year

The scale of what happened in the past twelve months is easy to undercount if you only see it in pieces. Put together, it represents one of the most significant recent declines in international enrollment across US higher education.

The Institute of International Education’s Fall 2025 Snapshot, which surveyed more than 800 colleges, found new international student enrollment down 17% compared with fall 2024. NAFSA’s economic analysis of that same data found it translated into a 7% decline in total international enrollment once students on Optional Practical Training are excluded, more than $1.1 billion in lost revenue, and nearly 23,000 fewer jobs supported nationally. The damage was not evenly spread. Undergraduate enrollment actually rose 2%, likely reflecting decisions and visa appointments made before the sharpest policy changes hit. Graduate enrollment fell 12%. Non-degree enrollment fell 16%, cutting into the student populations that often generate the greatest tuition revenue for institutions.

Common App’s own data, independent of IIE’s survey, shows the same direction. As of March 1 this year, international students submitting applications for the 2026-27 cycle were down 9% from the same point in the prior cycle.

The picture got worse heading into this spring. A May survey of 149 institutions, conducted by NAFSA and partner organizations, found 62% reporting lower international enrollment across undergraduate and graduate programs combined, with new undergraduate enrollment down 20% on average and new graduate enrollment down 24%. And the most current federal data, tracking actual international student arrivals into the country, shows the decline has not leveled off: arrivals were down 5% in March, nearly 8% in April, and 1% in May of this year compared with the same months last year, on top of a nearly 22% drop in arrivals last summer compared with the summer before.

In other words, this is not a single bad semester that colleges are already recovering from. It is a trend that has now compounded across multiple enrollment cycles, and it is still moving as your current clients build their lists for this fall. Ten days ago, the Department of Homeland Security published a final rule ending “duration of status,” the decades-old system that let international students remain in the country for as long as they made normal progress toward their degree. Starting September 15, students and scholars will instead be admitted for a fixed period tied to their program, capped at four years, and required to file a formal extension application with U.S. Citizenship and Immigration Services if they need more time. The rule represents another significant policy change affecting future international students. It is still being written.

Why this becomes your clients’ financial aid story

The instinct is to read international enrollment as a story about global politics, campus diversity, or immigration policy, all real, but distant from the domestic families an IEC actually advises. Follow the money instead, and the distance disappears.

International students are about 6% of enrollment nationally but generate roughly 12% of tuition revenue, according to research using 2016 enrollment and finance data, a share that may be higher today, though updated national estimates are not yet available. At institutions most dependent on this population, international students can account for more than 30% of total revenue. The reason is straightforward: nationally, only 16% of domestic undergraduates now pay the full sticker price, after years of colleges deepening their discounts to compete for a shrinking pool of American high school graduates. More than 80% of international undergraduates pay full price. Several public research universities, including Minnesota, Arizona State, Iowa, Purdue, Illinois, and Ohio State, charge international students between $874 and $5,218 more per year than even out-of-state domestic students, according to American Council on Education data. At UC Santa Barbara, international students pay more than three times what California residents pay, higher nonresident tuition that institutions use to support their overall finances, including institutional aid.

This is the mechanism Moody’s Ratings has been warning bond investors about since last summer. In its stress-test analysis of the colleges and universities it rates, Moody’s found that institutions where international students make up more than 20% of the student body, about 11% of the institutions it rates, face the sharpest financial exposure to enrollment declines. Modeling a relatively modest 10% drop in international enrollment, Moody’s found that 54 of the 392 institutions it rates would see a meaningful hit, at least half a percentage point, to their operating performance. Some reported enrollment declines this year exceeded the magnitude of the scenario Moody's modeled.

What this means in practice is that the international students your clients’ target schools have been losing were not simply filling seats. At a meaningful number of institutions, they provided unrestricted tuition revenue that many institutions used to support operations, academic programs, and institutional financial aid. When that revenue disappears, it does not vanish quietly into a spreadsheet. It shows up as a program that gets consolidated, a position that gets cut, or a tuition increase that lands on the family sitting across from you at the intake meeting.

The programs, positions, and price increases already gone

This is no longer a forecast. CollegeCuts has tracked more than 300 publicly reported instances of colleges eliminating programs, closing departments, or reducing staff since January 2024.

Beyond UNT, DePaul University laid off 114 employees after its international enrollment fell 30% overall, driven by a decline of roughly two-thirds in new international graduate students. The University of Southern California cut nearly 1,000 jobs, including academic advisers, after citing an expected decline in international students, whose graduate applications to USC had already fallen 23%, among its financial problems. Northwestern eliminated 425 positions, froze hiring, and delayed building projects, citing federal research funding freezes alongside projected international enrollment drops. The University of Texas at Arlington saw international enrollment fall 20% year over year, its steepest annual decline in a decade, with graduate international enrollment down 36.8% specifically. A draft UT-Arlington budget projected that decline in graduate students alone would cost the university between $13 million and $15.6 million in tuition revenue this fiscal year.

Syracuse University posted a rare budget deficit tied in part to declining international enrollment and raised tuition nearly 4% for the coming year, to a total cost of attendance of about $96,000. The New School, facing its own deficits and layoffs, raised tuition 3.5%, to roughly $93,000. The California College of the Arts announced it will wind down operations and close by 2027, with Vanderbilt University acquiring its campus, a decision its president tied in part to persistent budget pressures that international enrollment drops worsened. Boston University, Niagara College, and the Stevens Institute of Technology have each pointed to falling international graduate enrollment among the causes of their own budget cuts and layoffs.

None of this shows up on a campus tour or an admissions website. A family typically encounters it in October, in an acceptance letter for a version of a program that looked different when the student applied to it than the one an older sibling received two years earlier at the same school. What follows is how to check any school on a client’s list for this exact exposure before this fall’s applications go in, and what to actually do with what you find.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Issue #50. What began as an experiment has become a trusted weekly resource for IECs, admissions professionals, counselors, and higher education leaders around the world. Thank you for being part of the first fifty. Here’s to the next fifty.

For six weeks, a satirical youth movement has held the center of New Delhi, and a prominent Indian engineer, educator, and climate activist spent 26 days on hunger strike demanding an education minister’s resignation. On its face, that is not a story for a newsletter about American college admissions.

We’re covering it in full anyway, because the specific failure behind it, a single exam deciding a young person’s entire future, and that exam breaking under public scrutiny, is not unique to India this year. Nigeria and South Korea are living through their own versions of the same failure. Seeing all three side by side is the fastest way to understand exactly what a multi-factor American application protects your students from, and where it still doesn’t.

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What actually happened with India’s 2026 NEET paper leak and, separately, the CBSE marking scandal, and how the two fed into the largest youth protest movement India has seen in years

• Why Nigeria’s JAMB collapse and South Korea’s Suneung fight are not separate stories, but the same structural failure surfacing under a different flag

• What these episodes share underneath the local politics, and why the pattern matters more than any single country’s headlines

• Our take on how these systems should actually be redesigned, using India’s JEE process alongside NEET as a second, equally stark illustration of the same flaw

For paid subscribers:

• A framework for sizing single-point-of-failure risk in any exam system, so a worried family gets more than sympathy from you

• Specific talking points for Indian, Nigerian, and Korean client families right now, without overselling the US system to get there

• Where the actual practice opportunity sits this cycle for IECs serving pre-med and STEM-bound families reconsidering their options

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

What actually happened in Delhi

India’s National Eligibility cum Entrance Test, or NEET, is the single exam that decides medical school admission for the roughly 2.3 million students who register for it each year. There is no essay, no interview, and no alternate path around it. The exam went ahead as scheduled on May 3. Within days, the Central Bureau of Investigation confirmed that Chemistry, Biology, and Physics questions had circulated before students ever sat down, and it went on to arrest more than a dozen people, including a retired chemistry teacher tied to the exam’s Marathi-translation process and a doctor in Maharashtra accused of feeding leaked questions to students. The scale of the breach left the National Testing Agency no option but to cancel the exam entirely and reschedule it for June 21, this time on paper rather than digitally, with fees refunded and no new registration required.

A separate scandal was unfolding at the same time, in a different exam entirely. The CBSE, which runs India’s Class 12 school-leaving boards, had just introduced digital, on-screen marking for roughly 2 million students, and within weeks students were posting scanned answer sheets online that did not match their own handwriting. The uproar was serious enough that the government transferred the board’s chairman and secretary out of their posts. Two different exams, run by two different bodies, both broke within the same exam season. To a teenager watching the news that spring, whether the specific failure was NEET or CBSE mattered less than the pattern underneath it: the machinery meant to certify their future had stopped being trustworthy.

Both scandals fed into what has become the largest sustained youth protest India has seen in years, though the protest itself did not start as a protest about exams. On May 15, India’s Chief Justice, Surya Kant, compared unemployed young people to cockroaches during a Supreme Court hearing. The next day, a 30-year-old political communications strategist named Abhijeet Dipke, a Boston University graduate who had previously worked for the Aam Aadmi Party, launched a satirical online movement called the Cockroach Janta Party in response. What began as a joke absorbed the NEET and CBSE anger within weeks. On June 6, the Cockroach Janta Party held its first demonstration at Jantar Mantar in New Delhi (some of the images above are from that day), alongside established student groups including the Students Federation of India and the All India Students Association, demanding the resignation of Union Education Minister Dharmendra Pradhan. The movement’s founder has publicly put the number of students who died by suicide in the aftermath of the leak at more than twenty, a figure that has not been independently verified but that anchors one of the protesters’ core demands: compensation for those families.

Ladakhi engineer and education reformer Sonam Wangchuk, who had no direct connection to the NEET or CBSE issue, joined the protest in solidarity on June 28 and began an indefinite hunger strike. 20 days later, with his health deteriorating, Delhi police forcibly removed him from the protest site and hospitalized him against his and his family’s wishes. His wife called it an illegal detention and petitioned the Delhi High Court for his release. The court declined, though it later ordered the government to provide him with daily medical monitoring. On July 20, the first day of Parliament’s monsoon session, protesters attempted a march toward Parliament that police met with tear gas and baton charges. Police put the injury count, among both protesters and security personnel, at more than 170, with more than 100 protesters detained.

This is not the first time an Indian education minister has faced this exact accusation. On July 22, 2024, after a separate NEET paper leak forced a partial resit, Pradhan told the Lok Sabha there was no evidence of a leak in the National Testing Agency’s previous seven years of operation. Two years later, the agency canceled an entire exam cycle for the same reason. For a system with exactly one gate, a repeat failure of the same kind is not a data point. It is the argument for reform, made twice.

The same failure, three more flags

Read the Delhi headlines by themselves and this looks like an Indian political story. Widen the frame to the rest of 2025 and 2026, and it looks like a pattern.

Nigeria lived through its own version of it last year. The country’s Unified Tertiary Matriculation Examination, JAMB’s single computer-based test that determines university admission for the roughly 2 million Nigerians who sit it annually, released results in May 2025 showing more than three-quarters of candidates scoring below half marks. It turned out the problem wasn’t the students. A flawed server update from one of JAMB’s technical vendors had failed to properly upload candidate responses at 157 exam centers across Lagos and five southeastern states, corrupting the results of 379,997 people. Registrar Ishaq Oloyede apologized publicly and tearfully, later admitting his first instinct had been to resign before advisors talked him out of it, warning that stepping down would look like abandoning students at the exact moment they needed him most. 19 year old Faith Opesusi did not live to see the correction. She died by suicide after seeing her score, days before JAMB acknowledged the technical error publicly; a provisional admission offer to study microbiology, the course she had wanted, arrived in her family’s inbox roughly thirty minutes after her death was confirmed. Lawmakers from the affected region demanded Oloyede’s resignation, calling the episode a catastrophic institutional failure. He stayed on. JAMB ran a resit at no cost to affected candidates, though by the board’s own account, more than 5% of those called back didn’t show up for it either.

South Korea’s failure looks nothing like Nigeria’s or India’s on the surface, and that is exactly what makes it useful. The College Scholastic Ability Test, known as Suneung, is an 8 hour exam that more than half a million South Korean students sit each November, functioning as close to a single gate as any admissions system in the world. There was no leak and no server error. There was, instead, a design failure severe enough that Oh Seung-geol, the official who led the exam’s question-setting committee, resigned to take responsibility after the English section’s difficulty triggered national outrage, with roughly 370 formal objections filed against a single question. The government has spent 2 consecutive years trying to strip so-called killer questions, ultra-difficult problems that push students toward private tutoring for any real chance of answering them, and critics say the redesigned test still rewards students who can afford that tutoring. A resignation fixed the accountability problem. It did not fix the exam.

And in Bangladesh, the direction these movements can travel if left unresolved is not hypothetical. A protest that began in June 2024 over a government job quota, not an academic exam, escalated within weeks into a national uprising that killed hundreds of people, by the interim government’s own later estimate more than a thousand, and ended the 15 year rule of a sitting prime minister, who resigned and left the country. The grievance was different: access to public employment rather than admission to a university. The shape of the movement was the one now visible at Jantar Mantar. A generation concluded that the system deciding their future had stopped being fair, and made that conclusion loudly enough that the people running the system had to answer for it, whether they wanted to or not.

The pattern underneath the politics

Four countries, four governments, four different situations. What connects them is not ideology. It is architecture. Each is a system built around one high-stakes gate, run by a single centralized authority, with no meaningful alternate path and no independent body positioned to catch a failure before it reaches millions of students at once. When that kind of system works, it is efficient and, on paper, meritocratic. When it fails, it fails for everyone behind the gate at the same time, and there is no second door to try.

That architecture is not unique to the developing world, and the temptation to read this section as evidence that the American system is simply better should be resisted. Below is our own view on what a better-designed system would actually look like.

Our take: no single input should ever get to decide everything

Everything in this issue points to the same underlying design flaw, and it is worth stating directly rather than leaving it as subtext. A government exam system should never place the whole of a student’s future on the outcome of one test, administered by one body, on one day. NEET is the clearest current example, but it is not India’s only one. JEE, the exam chain that determines undergraduate admission to the IITs, runs on the identical architecture. More than 1.5 million students sat for JEE Main in 2026, of whom 250,182 qualified for JEE Advanced, of whom 56,880 passed, competing for just over 18,000 seats across all 23 IITs. A student’s four years of secondary coursework enter that process only as a minimum eligibility bar, 75% aggregate marks or a top 20-percentile finish, rather than as a genuine input into the decision itself.

We think education policy should be built from the ground up rather than administered from a single point at the top. That starts with weighing a student’s sustained classroom performance across multiple years alongside any standardized score, rather than reducing years of coursework to a pass or fail threshold. It also means distributing oversight of exam administration across more than one body, so a single failure cannot silently reach millions of students before anyone catches it, and building a remedy process that exists before a crisis forces one into being, rather than after.

None of that is a call to import the American model wholesale. A country moving away from single-exam admissions has to guard just as carefully against replacing one kind of unfairness with another. American admissions carries its own well-documented tilt toward students whose families can pay for essay coaching, curated extracurricular records, and the guidance that makes a multi-factor application legible to an admissions office. A well-rounded system only serves everyone if the factors that make it well-rounded are genuinely accessible regardless of what a family can spend, not just genuinely numerous.

The specific reform matters less than the principle underneath it. A system that asks one exam, one agency, and one day to carry the entire weight of a student’s future will eventually fail that student, and when it does, it will fail millions of them at the exact same moment. A system built on distributed, broad-based inputs fails more often, in smaller, more survivable ways. That tradeoff, gradual and repairable against rare and catastrophic, is the actual design choice every government running one of these systems is making, whether it has said so or not.

If your caseload includes families from India, Nigeria, or South Korea this year, or students from any country where a single test decides everything, some of them are already asking what these headlines mean for their own applications. Below is a framework for sizing that risk honestly, the specific things worth saying and not saying to a worried family, and where the real practice opportunity sits this cycle.

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Upgrade to Paid

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• The three colleges that have publicly confirmed AI touches how their applications get read, and what each system actually does

• Where the “half of all colleges” figure came from, and why a vice provost who was surveyed for it says it’s wrong

• What this changes about the file-reading process this newsletter walked through in Post 44 (linked below)

For paid subscribers:

• What 150,000 real admissions essays, compared directly against AI-generated ones by Cornell researchers, revealed about who uses AI and what happens to their applications

• Caltech’s written AI policy, including the specific line between accepted help and an Honor Code violation

• The exact script for the conversation with a family that wants a fast, polished draft and considers AI the way to get one

• How to check what a specific target school has actually said about this, since no Common Data Set field covers it and guessing is worse than silence

Share

The three confirmed cases

1. UNC Chapel Hill has quietly used an automated essay-scoring tool called PEG since 2019, something most families on its applicant list have never heard of. The Daily Tar Heel found out through a records request, obtaining the university’s contract with Measurement Incorporated, a Durham company, and then confirming the details with UNC’s own admissions leadership across a multi-part investigation that ran over several weeks. PEG scores essays on a one-to-four scale across six dimensions of writing mechanics: word choice, sentence structure, sentence variability, vocabulary, grammar, punctuation, and length. Rachelle Feldman, UNC’s vice provost for enrollment, told the paper the point was practical rather than punitive, freeing human evaluators to focus on content instead of mechanics while making scoring more consistent from one reader to the next. The university has spent close to $200,000 on the tool so far, under half a percent of the admissions office’s total budget, according to Feldman. Once the story drew public criticism, UNC updated its own admissions site with a direct statement: the university “uses AI programs to provide data points about students’ common application essay and their school transcripts”, while every application “is evaluated comprehensively by extensively trained human application evaluators.” So the AI score functions as one input into a human decision, not a replacement for one, at least according to UNC’s own account of how the system works. One more update worth noting for anyone tracking UNC specifically: starting with the 2026-27 cycle, the university eliminated its short-answer supplemental essays entirely, a separate move aimed at reducing applicant friction that doesn’t touch how the main application and transcript get evaluated.

2. Virginia Tech took a different route, and unlike UNC’s case, this one didn’t surface through a records request. The university announced it directly through its own news office. Starting with the 2025-26 cycle, Virginia Tech replaced its previous two-human-reader model with one human reader paired against one AI reader for every essay, building the underlying model in-house over three years rather than licensing it from an outside vendor. Juan Espinoza, the vice provost for enrollment management, laid out the safeguard in the university’s own announcement: “If the human and AI scores differ by more than two points, a second human reviewer is brought in to ensure consistency and fairness”. His explanation for why the change happened at all is fairly plain. Application volume has nearly tripled over two decades, and reading every essay twice by hand kept pushing notification dates later with each passing cycle, so the goal now is simply returning decisions to students more quickly even as the applicant pool keeps growing. Espinoza has also mentioned that peer institutions have reached out, curious but cautious, more interested in watching how Virginia Tech’s experiment plays out than in building something similar themselves just yet.

3. Caltech’s approach is the odd one out, since it doesn’t touch essays at all. Any applicant who submits a research paper or publication has to complete what the school calls a Viva Exam, built by a third-party company called InitialView that Caltech’s own site is careful to describe as “not affiliated with Caltech”. According to both Caltech and InitialView, the system generates follow-up questions specific to whatever research the student submitted, the student answers them on video, and a human reviews the recording afterward. Unlike the other two systems, this one isn’t evaluating writing quality or personality at all. It exists for a narrower purpose: confirming that the submitted research genuinely belongs to the applicant who claims it.

Beyond these three, the picture gets murkier fast. Other schools may well be doing something similar without having said so, and a fair amount of what circulates online about specific schools traces back to marketing content published by companies that sell AI detection or essay coaching services rather than by the schools themselves. A useful rule for evaluating any claim like that: if it doesn’t come with an on-record official or a document the school itself has confirmed, treat it as unverified rather than repeating it to a family as settled fact.

Where the number actually comes from

That statistic traces back to a single 2023 survey run by Intelligent, a for-profit education marketing site, drawing on fewer than 400 respondents, a fifth of whom weren’t admissions officers at all. Andy Borst, the University of Georgia’s vice provost for enrollment management, happened to be one of the people surveyed, and he later disputed the resulting claim in direct terms: “There is a zero percent chance admissions offices are using AI to read files in the way this internet survey and article imply”. His objection wasn’t that AI is absent from higher education generally. It was aimed specifically at the implication that admissions offices were already using AI to read application files, a much narrower and much less supported claim than the survey’s headline suggested.

The distinction matters more than it might seem. A statistic built on a shallow marketing survey gets repeated as settled fact anyway, article after article, precisely because it sounds specific and a little alarming, and because almost nobody checks the methodology before hitting publish. The three cases described above are more credible, due to the source of information from named people, on the record, describing systems that genuinely exist and explaining, in their own words, why they built them.

What this changes, and doesn’t

Post 44 (linked below) walked through how a human reader actually moves through a file: fast, structured, pattern-matching against thousands of prior applications, with real attention on the transcript and the counselor letter, and a closer read on the personal statement specifically because it’s the one part written entirely in the student’s own words. None of that has changed. What’s different, at these three schools and possibly others that simply haven’t said so publicly, is that parts of an application may now pass through an added layer of automated evaluation before or alongside the human one. Whether heavily AI-assisted essays tend to converge toward a more generic style, regardless of what any particular school’s system happens to be built to catch, is a separate question, and it’s the one the Cornell research below actually addresses.

What a family should do with all of this, given what the research says happens when a student leans on AI to write the essay itself, and what one top-tier school has now put in writing about where the line sits, is in the paid section below.

Upgrade to Paid

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Read more

Share

This newsletter grows from your shares. Send it to a fellow colleague today!

In today’s issue (free analysis):

• What the new DHS rule actually changes for F-1 and J-1 students, from the fixed four-year admission period down to the specific restriction on switching majors or transferring schools

• Why DHS says the change is overdue, what the government’s own data shows about how long students actually take to finish, and what nearly 34,800 public commenters said back

• The history behind this rule, including the nearly identical version proposed in 2020 and withdrawn a year later

For paid subscribers:

• The intake question most domestic-focused IECs are not asking, and the specific client profile that makes this rule relevant to a caseload with zero international applicants

• The financial mechanism, documented by economists and the American Council on Education, that connects a decline in international enrollment to the tuition and aid your domestic clients actually see

• What has actually been measured so far, what DHS itself admits it cannot yet quantify, and the two things that could still delay the September 15 effective date

• A concrete audit of program types, architecture, engineering co-op tracks, and combined-degree programs, where even an ordinary international applicant can run past the new four-year limit

Upgrade to Paid

Paid subscribers get the full analysis every issue, including the practitioner frameworks in the section below the paywall. If this newsletter helps you stay ahead in your work, a paid subscription is one of the better professional investments you can make. Many readers expense it as professional development.

Upgrade to paid

Already a paid subscriber? Thank you - you’re the reason this exists.

What actually changed

On July 17, DHS published a final rule in the Federal Register ending “duration of status,” the policy that has governed F-1 and J-1 visa holders since the early 1990s. Under duration of status, a student’s authorized stay was tied to enrollment itself. As long as a student remained enrolled and made normal progress, the government did not require a fixed departure date. That framework is gone as of September 15.

In its place, DHS is issuing a fixed admission period tied to the length of the program listed on a student’s Form I-20, capped at 4 years regardless of how long the actual program takes, plus a 30-day period built in for arrival and a separate 30-day period for departure. A student who needs more time, to finish a longer program, to begin a new one, or to start Optional Practical Training after graduation, must file a formal extension of stay with U.S. Citizenship and Immigration Services before the clock runs out. That filing requires biometrics and a fee, and it is a meaningfully different process than simply staying enrolled. The post-completion grace period for F-1 students, the window a student has to leave the country or change status after finishing a program, is also being cut from 60 days to 30 days.

The rule adds two restrictions that will matter more to how IECs think about program selection than the fixed clock itself.

1. An undergraduate F-1 student cannot change majors, change programs, or transfer to a different school during the first academic year, except through a narrow exception SEVP grants for things like a school closure or a natural disaster.

2. A graduate F-1 student faces a stricter version of the same rule for the entire length of their program, not just the first year: no changing programs, no changing majors, and no transferring schools without a similar SEVP exception.

Neither group can use F-1 status to pursue a second program at the same or lower level after completing one. English language training is separately capped at 24 months in total.

DHS itself estimates the rule will touch more than 2 million nonimmigrants a year and projects annualized compliance costs of $390 million to $392 million over the next decade, a figure that includes the added burden on designated school officials who now have to track fixed dates and file extension paperwork for students who previously needed neither.

Why DHS says this is necessary, and what the record shows

DHS’s stated justification is that open-ended status made it harder to verify, on an ongoing basis, that a student in the country was still doing what their visa said they would do. The agency points to cases of what it calls “forever students,” individuals who remained enrolled in language training or repeated coursework for years, in some instances more than two decades, without a fixed endpoint that would force a status review. Aligning F, J, and I categories with other nonimmigrant visa types, most of which already carry a fixed admission date, is the framing DHS uses throughout the rule.

The government’s own data complicates the case for a hard four-year line. In the rule itself, DHS states that international students typically earn a bachelor’s degree within four years of entering the country, and that 79% of international students are enrolled in either a four-year bachelor’s program or a two-year master’s program, the two categories the four-year cap was seemingly built around. But that argument rests on nominal program length, not actual completion time, and the government’s own separate data on domestic students undercuts the assumption that a four-year bachelor’s program reliably finishes in four years. The National Center for Education Statistics reports that only 44% of first-time bachelor’s degree recipients complete their degree within 48 months of first enrolling. A change of major, a switch in schools, a reduced course load during a difficult semester, any of the ordinary reasons a domestic student’s timeline stretches, applies just as easily to an international student on the same nominally four-year track, except that student now also has an immigration deadline attached to it.

The history behind this rule

This is not a new fight. DHS proposed a nearly identical rule in 2020, during the first Trump administration, and withdrew it in 2021 before it took effect. The current version was proposed on August 28, 2025, and drew more than 34,800 public comments before the comment period closed on September 29, the large majority opposed. A comment letter organized by the American Council on Education and signed by more than 55 higher education associations argued the rule would strain advising offices and discourage enrollment, while addressing a fraud problem the existing SEVIS tracking system was already built to catch. Fanta Aw, NAFSA’s executive director and CEO, has called the underlying justification “a solution in search of a problem”. DHS itself, in the final rule’s own preamble, concedes that international enrollment could decline as a result, while stating that it lacks a comparable precedent to estimate by how much.

No lawsuit has been filed against the final rule as of this writing, though immigration law firms tracking the rule expect one on administrative law grounds once it takes effect, and the rule’s September 15 date could still be affected by action under the Congressional Review Act. For now, the current regulations remain fully in force, and nothing changes until the effective date arrives.

The rule’s direct reach stops at the students who hold F-1 or J-1 status. For a boutique IEC whose caseload is entirely domestic, that might read as a rule to skim past. It is worth one more look before you do.

Upgrade to Paid

Higher Ed Insights is reader-supported. If you are reading this as a free subscriber and you work in college counseling, financial aid, enrollment management, or high school education, the paid section is written directly for you.

Read more