The 2025–26 Common Data Sets Are Landing: New Early Decision Numbers From 32 Colleges, and the Nine We Can Compare

Updated October 9, 2026: seven more 2025–26 filings added (Bowdoin, Chapman, Columbia, Northwestern, Penn, Smith and Virginia), with five new year-over-year comparisons.

Every autumn, colleges start posting a new Common Data Set, the standardized filing that reports, among much else, how many students applied Early Decision and how many were admitted. The 2025–26 filings are arriving now, and as of early October, 32 colleges in our catalog have published Early Decision counts for the most recent cycle. This is the first look at those 2025–26 Common Data Set early decision numbers, and at the nine colleges where a filing from the year before lets us see what moved.

Who has filed so far

The 32 with 2025–26 Early Decision counts, each linked to its own page: Brown, Columbia, Cornell, Dartmouth, Northwestern, Penn, Vanderbilt, Rice, Johns Hopkins, Washington University in St. Louis, Emory, Carnegie Mellon, Tufts, Boston University, Case Western Reserve, Lehigh, Villanova, University of Rochester, Rochester Institute of Technology, Baylor, Santa Clara, Chapman, American, William & Mary, Virginia, Vermont, Wesleyan, Bowdoin, Haverford, Smith, Skidmore and Harvey Mudd.

The range inside that group is wide. Vermont admitted 91.7 percent of its early applicants; Johns Hopkins admitted 10.9 percent. A few colleges people ask about are still on their 2024–25 filing, including Boston College and Duke, whose 2025–26 filing leaves the Early Decision rows blank. Their pages say which year they are on.

The nine we can compare

For most colleges, this year's filing is the first one we hold with a full early round in it, so there is nothing to compare it to yet. Eleven colleges have two consecutive years on file. Chapman's 2025–26 filing repeats its 2024–25 Early Decision counts exactly, and we could not re-check Virginia's 2024–25 early counts against its filing, so both are left out below, which leaves nine.

Six of the nine early rates went up and three went down. Only one moved because the college admitted many more early applicants.

The pool moved more than the policy did

Look at the admits column. Brown admitted 907 early applicants this year and 898 the year before, nine more. William & Mary admitted eleven fewer. Wesleyan admitted twenty more, Smith eighteen more, and Columbia, Northwestern, Bowdoin and Penn between 24 and 31 more. Those are the kinds of differences you would expect from one year to the next at a college that has settled on how big its early round should be.

Now look at the applicants column. Brown's early pool shrank by 19 percent, from 6,251 to 5,062. William & Mary's shrank by 7 percent, Wesleyan's by 7 percent. The rate rose at all three because the denominator fell, not because the colleges opened the door wider. If you had applied early to Brown in either year, you were competing for roughly the same 900 places; this year there were 1,189 fewer people competing for them.

It runs the other way too. Northwestern's early pool grew by 15 percent, Penn's by 9 and Smith's by 20, while their early admits rose only 2 to 5 percent, and all three early rates fell: Northwestern from 23.0 to 20.4 percent, Penn from 14.2 to 13.3, Smith from 38.2 to 33.5. Columbia's and Bowdoin's pools barely moved, and their rates rose by 0.8 and 1.4 points on about two dozen more admits each.

Cornell is the exception, and it is the one we have written about on its own: its early pool was flat and its early admits rose by 728. That is a change in what the college did, not in who showed up.

Why this matters for the round that is open now

The reading is simple. At most colleges, the number of early admits is close to a fixed quantity. The early rate you see is that quantity divided by however many people chose to apply early that year. Which means the rate you are looking at is partly a measurement of other applicants' decisions, and those decisions are about to shift.

This cycle, the University of Florida, Florida State, Michigan and USC all run binding early rounds that did not exist a year or two ago. Every student who uses one of them is a student who is not using Early Decision somewhere else. Where those students would otherwise have applied, the early pool gets a little smaller, and if the college admits its usual number, the rate drifts up. Nobody can say by how much. But a 2025–26 number is the most recent reading of a pool that is already changing, not a fixed property of the college.

How to hold a one-year number

Three habits, in order of how often they matter:

1. Check which year the number is from. A 2024–25 filing describes applicants who heard back in December 2024. Our pages print the year next to every figure for exactly this reason.
2. Look at the counts, not only the rate. A rate built on 396 applicants (Vermont) moves far more easily than one built on 10,057 (Cornell). One more admit at a small college is a quarter of a point.
3. Read the early rate next to the rate for everyone else. Brown's early rate rose 3.5 points this year; so did its regular rate, by 0.7 points. The gap between the two rounds, which is the thing an early applicant is actually buying, moved less than either number alone suggests.

The comparison table above is what two years of a college's own filing can tell you. What it can't tell you is where your own score sits in that college's range, which is the other half of an estimate. Start with the school you're leaning toward: what it published, and your estimate there, are free. The full answer for your whole list is the paid document.