A college's acceptance rate is printed as one number. Dartmouth, 6.0%. Penn, 5.4%. Columbia, 3.9%. Students build lists around those figures and never notice that each one is an average of two very different contests — the Early Decision round and everything after it — held months apart, with different rules and different odds.
The schools publish enough to separate them. Almost nobody does.
How much of the admitted class is settled in December
Every Common Data Set filing reports early decision applications and early decision admits alongside the first-year totals. Divide one by the other and you get a figure that never appears in a press release: the share of the whole admitted class that came through the binding round.
Dartmouth, 2025–26 filing: 3,552 early applicants out of 28,230 total — 12.6% of the pool. They received 768 of 1,699 offers. That is 45.2% of every admission Dartmouth made that year, awarded to one applicant in eight.
Penn, 2024–25 filing: 13.3% of applicants applied early; they took 35.1% of the offers.
Columbia, 2024–25: 10.0% of applicants, 34.2% of the offers.
Vanderbilt, 2025–26: 12.9% of applicants, 33.7% of the offers.
Brown, 2025–26: 11.8% of applicants, 33.5% of the offers.
And the extreme case. Northeastern's 2024–25 filing shows 3,466 early applicants in a pool of 98,425 — 3.5%. They received 1,492 of 5,133 admits: 29.1%. One applicant in twenty-nine, taking nearly three offers in ten.
Why the share of admits understates it
Those figures are shares of offers. The share of seats is larger, and the filings tell you why without stating it. Early decision admits are bound to enroll; regular admits are not, and at selective schools a large fraction of them go elsewhere. So when a school makes a third of its offers early, the students in those offers become considerably more than a third of the freshman class. The school knows this. It is the whole point of running a binding round.
What is left for everyone else
Subtract the early round from both totals and you get the rate that actually applied to a Regular Decision file.
At Dartmouth, 24,678 applicants competed for 931 remaining offers: 3.8%. The published 6.0% overstates the Regular Decision contest by more than half.
At Columbia, 2.8% for everyone else. At Penn, 4.0%. At Vanderbilt, 4.1%. At Northeastern — where the headline rate is 5.2% — the rate for everyone outside the early round was 3.8%.
If you are planning a Regular Decision application to a school like these, the rate for everyone else is your number. The headline figure was pulled upward by a round you chose not to enter.
Not every school works this way
The pattern is concentrated at the top of the selectivity range and fades quickly below it. Rochester Institute of Technology's 2025–26 filing shows 4.5% of applicants applying early and receiving 4.6% of offers — the early round is a service, not a strategy. Baylor: 1.3% of applicants, 1.9% of offers. Villanova sits between, with 6.5% of applicants taking 13.8% of offers.
For schools like RIT and Baylor, the headline acceptance rate is an honest description of any round. For Dartmouth, Penn, or Columbia, it is a blend that describes neither.
What this changes
Two things.
First, if a school fills a third or more of its admitted class early, the cost of not applying early is larger than any early-round bonus. You are not merely giving up an advantage; you are entering the contest after a third of the prizes have been handed out, with the remaining rate well below the number on the brochure.
Second, this does not mean everyone should apply early somewhere. The early pool at these schools is not a random sample. It is thick with applicants already inside the admitted range, plus recruited groups. The 45.2% at Dartmouth describes the round, not any particular applicant in it. What the share tells you is where the school makes its decisions. Whether you belong in that round is a separate question, answered by the school's admitted range and your place in it.
Doing this for your own list
The arithmetic takes a minute per school. Find the school's Common Data Set, which most post under institutional research, and open section C. Line C1 gives total applicants and total admits. Line C21 gives early decision applicants and admits. Divide early admits by total admits for the share of offers. Subtract both early figures from both totals and divide for the rate everyone else faced. Schools that leave C21 blank are telling you something too; a blank is not the same as a small number, and it is worth knowing which one you are looking at.
Where the numbers come from
Early decision applications and admits are from each school's Common Data Set, section C21; total applicants and admits are from section C1 of the same filing; the cycle is named beside each school. The rate for everyone else is derived by subtraction. Headline acceptance rates are the same filings' C1 totals. None of it is estimated, and all of it is public.
PrepToDone reads these filings back against your actual list — the early rate, the rate for everyone else, and where your score sits — free, without an account. Deciding which school gets the application is the paid part: $49, once. Put your list in.