We Checked the Early Decision Section of 17 Schools' Common Data Sets. Ten Were Wrong.

While building a series decomposing published acceptance rates school by school, we ran a simple check on every college whose Common Data Set includes Early Decision numbers: does Section C21, the Early Decision section, describe the same cohort as the rest of the filing it sits in? At 10 of the 17 schools where the check could be run, the answer was no. Section C21 was stamped with the prior year's cohort label inside a current filing, and the figures it reported didn't match what that prior year's own filing said about the same cohort.

The clearest case is Howard University, where the problem shows up without even needing an old filing for comparison. Howard's two most recent filings both label their Early Decision section "Fall 2024," and they report different numbers: 348 applicants and 172 admits in one filing, 430 applicants and 363 admits in the other. Brandeis's current filing lists Fall 2024 Early Decision as 680 applicants and 242 admits, while the actual Fall 2024 filing, filed a year earlier, put it at 765 and 323. Bates's current filing says 949 and 318 for Fall 2024; Bates's own Fall 2024 filing said 1,128 and 306. Carleton shows 760 and 263 against its own prior filing's 667 and 244. Harvey Mudd shows 594 and 109 against 656 and 106.

This matters because Early Decision rates are among the most heavily cited statistics in admissions advice. "Applying ED roughly triples your odds," or some version of that claim, shows up constantly, and it's almost always built on a Common Data Set figure taken at face value. At most of the schools we checked, that figure describes a different admissions cycle than the filing claims.

What the Common Data Set Is, and Why Section C21 Matters

The Common Data Set (CDS) is a standardized survey most four-year U.S. colleges fill out once a year and publish, usually as a PDF on an institutional research office's website. It exists so colleges, guidebook publishers, and ranking organizations can all ask for the same information in the same format, instead of every school reporting admissions data its own way. Section C1 covers overall applications, admits, and enrolled students for the entering class. Section C21 asks a narrower question: of the applicants who applied under a binding Early Decision plan, how many applied and how many were admitted, broken out from the Regular Decision pool.

C21 is the only place most schools publish an Early Decision figure at all. Admissions offices rarely put ED counts anywhere else — not in press releases, not in class profiles, not on their websites. When a guidebook, ranking site, or admissions blog cites an "ED acceptance rate," the number underneath it almost always traces back to this one section of the CDS. If that section is mislabeled, everything built on top of it inherits the error.

The Ten Filings

Here is what each school's current filing reports for Early Decision, set against what the prior year's own filing reported for the same stated cohort. All figures below come directly from each school's own Common Data Set filings.

For RPI, Rochester, and Miami, we confirmed the mislabeling and mismatch without extracting exact counts from both filings; for the other six, the numbers above show the disagreement directly. In every case, the underlying issue is the same: the section labeled with a given year doesn't match what that year's own filing reported.

Seventeen is the number of schools where Early Decision data existed and the year-over-year check could actually be run — it is not a claim about how the Common Data Set behaves industry-wide, and we did not sample colleges at random. Six schools passed cleanly: Northeastern, Cornell, UVA, Wake Forest, Drexel, and Claremont McKenna all had matching cohort labels and genuinely different year-over-year figures, which is what a correctly updated filing should look like.

Why an Internal Check Can't Catch This

A natural instinct when reading a Common Data Set is to sanity-check the numbers against each other: does Early Decision plus Regular Decision add up to the total in Section C1? At every school in this list, it does. That's exactly what makes the error hard to spot.

Many published Early Decision analyses, including some of our own earlier drafts in this series, derive a Regular Decision rate by subtracting the ED figures from the school's overall total: RD applicants equal total applicants minus ED applicants, RD admits equal total admits minus ED admits. That subtraction will always balance, because ED and non-ED are defined to sum to the total — it's arithmetic, not verification. If the ED figure being subtracted belongs to the wrong year, the resulting RD figure is also wrong, and it will still add up perfectly with everything else in the filing. Internal reconciliation cannot detect a cohort-label error. The only way to catch it is to look outside the filing, at what the prior year's filing said about the same class.

The Check You Can Run Yourself

This is genuinely something a reader can verify without special access, in about five minutes per school:

1. Open the school's current Common Data Set and find Section C21. Read the cohort year printed on that specific block — not the filing's cover year, the label inside C21 itself.
2. Find that school's Common Data Set from the year before, and look up its own C21 section for the same stated cohort.
3. Compare the two. If the applicant and admit counts match, the section is current. If they disagree, treat both as unusable until the school clarifies, and flag any rate built on either rather than publishing it as fact.
4. Where possible, look for an ED figure the school published somewhere outside the CDS — a counselor-facing profile, a fact book — and check whether it agrees. A figure confirmed by two independent sources is worth more than one from the CDS alone.

What This Does and Doesn't Mean

The likely explanation is mechanical, not deceptive: Section C1 gets updated with each year's numbers, and Section C21 is left carrying forward last year's block, probably because the form is built from the prior year's file and one section doesn't get refreshed. Nothing here indicates schools are concealing information or manipulating figures, and we're not suggesting either. It reads as a template error, the kind that happens when a long annual form gets copied forward and one section is missed.

It also doesn't mean the CDS itself is untrustworthy. It remains the best public source for admissions data available, and this series is built on it. The lesson is to read Section C21 carefully rather than at face value — check the year printed on that block against the prior filing before citing an Early Decision rate.

We've run into this in practice twice already. Our Howard University acceptance rate breakdown shows both of Howard's contradictory Early Decision figures side by side rather than picking one to publish as fact, because neither can be confirmed. Our Wake Forest acceptance rate breakdown covers a related problem: Wake Forest's ED data didn't get mislabeled, it stopped being published at all, so the last reliable breakdown on record is now several years old.

Where to Go From Here

If you're building a college list around Early Decision odds, don't take any school's ED rate at face value — check whether it's actually current. Our free college list builder helps you weigh schools using rates we've verified this way, not just headline numbers. Our application deadline tracker keeps ED, ED II, and RD deadlines straight across your list. And if you're deciding whether binding ED makes sense at a school where the public data is murky, a free strategy call is a good place to work through it with someone who has actually pulled the filings.