A “safety” with a 25 percent admit rate is not a safety. The reach, match, safety framework most parents and high-school counselors still use was built for an admissions environment that no longer exists. It has been broken for roughly a decade, and using it in 2026 is one of the most common reasons strong applicants end up with thin April outcomes.
We see it every cycle. A family arrives with a list of twelve schools, neatly tiered by a well-meaning counselor: three reaches, six matches, three safeties. The “safeties” are Boston University, Northeastern, and Tulane. The “matches” include Vanderbilt and Emory. On paper, the math looks comfortable. On April 1, two of the three “safeties” have rejected the student outright, and the family cannot understand what happened.
What happened is that the framework was wrong before the file was even opened.
What a “safety school” actually means now
The original reach, match, safety language comes from a moment, roughly 1995 to 2008, when most selective American universities admitted between 25 and 55 percent of applicants. In that environment, a school whose median admitted student looked weaker than your child was, by reasonable statistical logic, a high-probability admit. You could call it a safety and mean it.
That world is gone. The NACAC State of College Admission report has documented a sustained, structural decline in admit rates across the entire selective tier, not just the Ivies. The Common App’s annual reports show application volume up roughly 30 percent over the last five cycles at the schools families most want. The Common Data Set filings (Section C) from school after school tell the same story: admit rates that were 35 to 50 percent a decade ago are now 10 to 18 percent.
The replacement definition we use with families is narrow on purpose. A true safety in 2026 has three properties at once. First, a published admit rate above roughly 50 percent (ideally 65 percent or higher). Second, your child’s academic profile (GPA, rigor, test score if submitted) sits at or above the 75th percentile of the most recent admitted class per the Common Data Set. Third, the family is financially indifferent to the sticker price, or the school has guaranteed-aid formulas the family already qualifies for. If any one of those three fails, the school is not a safety. It is a target or a reach with optimistic framing.
By that definition, most lists we inherit from outside counselors contain zero true safeties. That is the open loop. We will come back to how to close it.
Yield protection, and why “overqualified” can equal “rejected”
The second reason the old framework fails is a phenomenon admissions officers rarely discuss on the record but every dean of enrollment manages around: yield protection, sometimes called Tufts syndrome after the school where the pattern was first widely observed.
Yield is the percentage of admitted students who enroll. It is a major input to US News rankings, to bond ratings, and to the internal politics of any admissions office. A school that admits a clearly overqualified applicant, one whose stats and trajectory suggest they will use the school as a backup for a higher-ranked peer, risks an admit that does not yield. Multiply that across a few thousand files and the school’s yield rate falls, its rank slips, and its discount rate climbs.
The rational institutional response is to defer or reject applicants who look like obvious flight risks, even when those applicants would clearly succeed academically. Tufts, Washington University in St. Louis, Boston University, and several others have been studied repeatedly in this context. Read their Common Data Sets alongside their published admit profiles and the pattern is visible: admitted-student stats that look surprisingly tight, suggesting the office is actively filtering for fit and likelihood-to-enroll, not just credentials.
For a HNW family, this is the cruelest part of the old framework. The very profile that makes your child competitive at Stanford, a 1560 SAT, a 4.0 unweighted, three APs in junior year, a serious extracurricular spike, makes them statistically suspicious at the schools they have been told are their safeties. Without a clear, documented signal of genuine interest, the school assumes you are using them as insurance and protects its yield by saying no.
IECA’s published guidance on school-list construction has been pointing at this for years: the list has to be built around fit signals the school can actually see, not just statistical comfort.
The replacement framework: four axes, scored per school
We retired reach, match, safety inside our office around 2018. What replaces it is a four-axis score we build per school, per student. Every school on a list has to clear a defensible threshold on all four. The axes are:
1. Institutional fit. Does the academic program, the social culture, the size, the location, and the post-graduate outcome profile actually serve this specific student? A student who wants a 12-person seminar on Heidegger in sophomore year does not belong at a 40,000-person research university, regardless of rank. Fit is not a feeling. It is curriculum, advising structure, lab access, and alumni outcomes in the student’s intended field.
2. Financial fit. For most of our families this is trivial (full pay, no aid filings). For others it is decisive. Some schools meet 100 percent of demonstrated need with no loans (Princeton, Yale, Harvard, MIT, a handful of others). Some are need-aware in admission. Some offer merit scholarships at scale (Vanderbilt, USC, Chicago through specific named awards). The financial-fit answer determines which schools belong on the list at all, and which ones the family should not waste an essay on.
3. Statistical reality. This is the honest read of admit probability, not the marketing version. We look at the most recent three years of Common Data Set filings, the school’s posted middle-50 ranges, the admit-rate trend line, and any known shifts (a new president, a strategic plan that changes class size, an athletic-recruiting reset). A student in the 60th percentile of admitted stats at a school with a 6 percent admit rate is not a “match.” They are a reach. Naming it correctly changes how the family makes decisions in March.
4. Demonstrated interest, scored. For schools that track it (and many publicly say they do not, while their Common Data Set Section C7 quietly says they do), we build a per-school interest plan: campus visit logged, info-session attendance, regional admissions officer contact, supplemental essay specificity, optional interview taken. We score each school 0 to 5 on the strength of the signal the file will carry by submission. A school below a 3 either gets a real interest investment or comes off the list.
A list built this way usually contains nine to twelve schools, not eighteen, and the family knows in October which ones are real.
Admit rates, then and now
| School | Admit rate, ~2008 | Admit rate, most recent CDS |
|---|---|---|
| Boston University | ~58% | ~11% |
| Tufts | ~27% | ~10% |
| Washington University in St. Louis | ~21% | ~12% |
| Northeastern | ~42% | ~6% |
| University of Chicago | ~35% | ~5% |
(Source: published Common Data Set filings and university fact-book archives. Rates rounded.)
A parent looking at their own college decision in 1998 has no useful intuition about any of these institutions today. The names are the same. The schools are not.
What to do next
If your child is entering the application cycle in the next 18 months, three actions will save the list.
First, pull the Common Data Set for every school currently on the list. Read Section C7 (admission factors) and Section C9 (admitted-student profile). If the school’s median admit looks stronger than your child, the school is not a safety regardless of what the counselor called it. If C7 lists “level of applicant’s interest” as “considered” or “important,” the school tracks demonstrated interest and the file needs to show it.
Second, replace any school that fails the four-axis test. The replacement is usually one of three categories: an honors college at a flagship state university where the student is clearly in the top quartile, a strong liberal arts college with a higher admit rate and a culture of yield protection working in the student’s favor (when interest is genuine), or an international option (St Andrews, Trinity College Dublin, McGill, Bocconi, HKU) where the academic match is real and the admit logic is transcript-driven rather than holistic-and-political.
Third, build the demonstrated-interest plan in writing, per school, by August before senior year. Visits booked. Regional officer identified. Supplement angles drafted. Interview windows on the calendar. A list without an interest plan is a list that will underperform in March, no matter how good the essays are.
The reach, match, safety framework felt safe because it was simple. The current environment is not simple. A list that reflects that reality, four axes, scored, with a real plan behind each name, is the only kind of list we send into a 2026 cycle.