Duke admitted more first-year students to the Fall 2023 entering class than to any other class in the past five years, and it did so in the one cycle when its applicant pool shrank. Applications fell from 49,523 for Fall 2021 to 46,366 for Fall 2023. Admits rose from 2,911 to 3,145. The admit rate climbed from 5.88 percent to 6.78 percent, the only increase in the period and the only reading above 6 percent in this window.
Then the pool came back. Fall 2024 brought 51,795 applications, Fall 2025 brought 55,541, and the admit rate fell to 5.24 percent. Through all of it the entering class barely moved: 1,744 students, then 1,742, then 1,740, then 1,749. Duke assembled the same class four times from four very different applicant pools.
That fixed denominator drives most of what follows. When a university holds its class at roughly 1,745 students and its applications swing by 9,000 in two years, the admit rate becomes a measure of demand rather than a measure of policy. The more useful questions are where the applications came from, which round they arrived in, and what the university charged the students who enrolled.
What Duke filed, and what it did not
Duke published no Common Data Set for 2022-23. The Fall 2022 entering class has no filing, and figures assembled from press releases or federal reporting do not follow CDS definitions closely enough to sit in the same row. The tables below mark that year as not filed rather than estimate it. Four filings carry the analysis: 2021-22, 2023-24, 2024-25, and 2025-26.
That is not the only gap. The 2025-26 filing leaves the early decision application and admit counts blank, omits the residency breakdown entirely, and reports no waiting list figures. The 2024-25 filing skips the SAT composite range while reporting the section scores that produce it. Duke discloses less through the Common Data Set now than it did in 2021, and a reader comparing Duke to a peer should confirm the row exists before drawing a conclusion from a blank.
One counting note governs everything here. Duke’s admissions office announced 58,712 applications and 2,790 admitted students for the Class of 2029. The 2025-26 filing reports 55,541 and 2,910 for that same class. The Common Data Set counts actionable applications, meaning completed files that received a decision, while a press release counts everything that arrived. Both figures are accurate. Every number in this report follows CDS definitions.
Applications, admits, and yield
| 2021-22 | 2022-23 | 2023-24 | 2024-25 | 2025-26 | |
| Entering class | Fall 2021 | not filed | Fall 2023 | Fall 2024 | Fall 2025 |
| Applications | 49,523 | n/a | 46,366 | 51,795 | 55,541 |
| Admitted | 2,911 | n/a | 3,145 | 2,957 | 2,910 |
| Admit rate | 5.88% | n/a | 6.78% | 5.71% | 5.24% |
| Enrolled | 1,744 | n/a | 1,742 | 1,740 | 1,749 |
| Yield | 59.9% | n/a | 55.4% | 58.8% | 60.1% |
| First-year retention | 98.0% | n/a | 98.0% | 98.0% | 97.0% |
Source: Duke University Common Data Set filings, sections C1 and B22. Entering classes of Fall 2021, Fall 2023, Fall 2024, and Fall 2025. No filing was published for 2022-23. Admit rate = students admitted divided by applications. Yield = students enrolled divided by students admitted.
Applications and admit rate, Duke University, four Common Data Set filings
Yield deserves as much attention as the admit rate. It fell to 55.4 percent for Fall 2023, the same cycle in which Duke admitted its largest group, and returned to 60.1 percent for Fall 2025. A university that converts three of every five admitted students has unusual control over its own class size, and Duke used that control to hold enrollment flat while the admitted group moved by more than 200 students.
Students admitted and yield rate, Duke University, four Common Data Set filings
Retention slipped in the most recent filing. Duke reported 98.0 percent first-year retention in three straight filings, then 97.0 percent for the Fall 2024 cohort measured in Fall 2025, with 1,687 of 1,740 students still enrolled. A single point is not a trend, and at this level the difference amounts to roughly 17 students. Six-year graduation rates have held near 96 percent throughout.
The early round carries a growing share of the class
| 2021-22 | 2022-23 | 2023-24 | 2024-25 | 2025-26 | |
| ED applications | 5,060 | n/a | 4,881 | 6,013 | not reported |
| ED admits | 828 | n/a | 962 | 1,042 | not reported |
| ED admit rate | 16.4% | n/a | 19.7% | 17.3% | n/a |
| ED admits as share of class | 47% | n/a | 55% | 60% | n/a |
| Approx. regular admit rate | 4.7% | n/a | 5.3% | 4.2% | n/a |
Source: Duke University Common Data Set filings, sections C1 and C21. ED admit rate = early decision admits divided by early decision applications. Share of class = early decision admits divided by total students enrolled. Approximate regular admit rate = (total admits less ED admits) divided by (total applications less ED applications); it is approximate because Duke’s reported ED admit count includes students admitted during the regular round. The 2025-26 filing left both ED counts blank.
Early decision applications and admit rate, Duke University
Duke defines its early decision admit count more broadly than its December announcement does. The 2023-24 filing states plainly that the 962 figure includes students admitted through the QuestBridge National College Match and students who applied early decision, were deferred, and were then admitted in the regular round. That single sentence resolves a discrepancy that confuses many applicants. For the Class of 2029 Duke announced 6,627 early applications and 849 admits, a rate of 12.8 percent, while the CDS-style rate for the prior class ran above 17 percent. Neither figure is wrong; they count different groups.
Even allowing for that breadth, the direction is unmistakable. Early decision admits equaled 47 percent of the entering class for Fall 2021 and 60 percent for Fall 2024. Applicants who wait for the regular round are competing for a shrinking remainder, and the approximate regular admit rate of 4.2 percent for Fall 2024 sits well under half the early rate. The choice of round is now among the largest levers an applicant controls.
Residency separates applicants more sharply than any essay
| Applications | Admits | Admit rate | Enrolled | |
| Fall 2023, North Carolina | 3,765 | 376 | 9.99% | 118 |
| Fall 2023, out of state | 32,191 | 2,370 | 7.36% | 1,429 |
| Fall 2023, international | 10,410 | 399 | 3.83% | 195 |
| Fall 2024, North Carolina | 4,748 | 377 | 7.94% | 267 |
| Fall 2024, out of state | 31,906 | 1,995 | 6.25% | 1,178 |
| Fall 2024, international | 15,141 | 585 | 3.86% | 295 |
Source: Duke University Common Data Set filings 2023-24 and 2024-25, section C1 residency table, reported by known physical address at time of application. Admit rate = admits divided by applications. The 2021-22 form did not collect this breakdown and the 2025-26 filing left it blank. Section F1 reports out-of-state share on a different basis and does not reconcile precisely with the enrolled counts above.
Admit rate by residency category, Duke University, Fall 2023 and Fall 2024 entering classes
International applicants faced roughly half the odds of domestic applicants in both years Duke reported the split, at 3.83 percent and 3.86 percent. The pool behind those rates grew 45 percent in a single cycle, from 10,410 to 15,141, and Duke expanded international admits almost proportionally, from 399 to 585. The rate held steady because the university let the admitted group grow with the pool rather than capping it.
North Carolina tells the opposite story. In-state applications rose 26 percent between the two cycles while in-state admits stayed flat at 376 and 377, compressing the resident admit rate from 9.99 percent to 7.94 percent. That compression arrived in the second year of Duke’s Carolinas tuition commitment, which is exactly what a successful recruitment program looks like from the inside: more applicants, the same number of seats, and a harder path for each individual applicant. In-state students still hold the best odds at Duke, but the advantage narrowed by two full percentage points in one year.
Testing: half the class submits, and the ACT is fading
| 2021-22 | 2022-23 | 2023-24 | 2024-25 | 2025-26 | |
| Submitted SAT | 47% (825) | n/a | 47% (824) | 48% (824) | 48% (843) |
| Submitted ACT | 46% (810) | n/a | 30% (519) | 30% (526) | 27% (473) |
| SAT composite, 25th to 75th | 1490-1560 | n/a | 1520-1570 | not reported | 1510-1570 |
| SAT math, 25th to 75th | 750-800 | n/a | 780-800 | 760-800 | 770-790 |
| ACT composite, 25th to 75th | 33-35 | n/a | 34-35 | 34-35 | 34-35 |
Source: Duke University Common Data Set filings, section C9. Percentages express submitters as a share of the enrolled first-year class; students submitting both tests appear in both rows. The 2024-25 filing omitted the SAT composite range while reporting section scores.
Share of enrolled first-year students submitting each test, Duke University
SAT submission has been remarkably stable at 47 to 48 percent of the enrolled class across every filing. ACT submission collapsed, from 46 percent for Fall 2021 to 27 percent for Fall 2025, a loss of 337 students in absolute terms while the class size held. Duke remained test-optional throughout the period, so roughly half the class has enrolled without submitting an SAT score in each of the last five years.
Duke’s own stated weighting shifted more than the scores did. In the 2021-22 filing the university rated standardized test scores as very important, alongside rigor of record, grade point average, essays, recommendations, extracurricular activities, talent, and character. By the 2023-24 filing test scores had dropped to considered, and they have stayed there. The application essay made the same fall, from very important in 2021-22 to considered in 2023-24 and 2024-25, then recovered to important in 2025-26. Class rank moved the other way, from unrated to important. Racial and ethnic status, rated considered in 2021-22, vanished from the form after the 2023 Supreme Court decision. Religious affiliation slid from considered to not considered.
Four factors never moved: rigor of secondary school record, academic grade point average, recommendations, and character and personal qualities all sat at very important in every filing, joined by extracurricular activities and talent. Duke’s evaluation has become more dependent on the transcript and the people who write about the student, and less dependent on anything the student produces on test day.
Who enrolls
| First-year students | 2021-22 | 2023-24 | 2024-25 | 2025-26 |
| Asian | 329 (18.9%) | 421 (24.2%) | 337 (19.4%) | 439 (25.1%) |
| White | 670 (38.4%) | 625 (35.9%) | 598 (34.4%) | 573 (32.8%) |
| Hispanic or Latino | 226 (13.0%) | 173 (9.9%) | 179 (10.3%) | 197 (11.3%) |
| Black or African American | 143 (8.2%) | 131 (7.5%) | 154 (8.9%) | 113 (6.5%) |
| Two or more races | 135 (7.7%) | 128 (7.3%) | 124 (7.1%) | 102 (5.8%) |
| International | 172 (9.9%) | 184 (10.6%) | 160 (9.2%) | 193 (11.0%) |
| Race or ethnicity unknown | 65 (3.7%) | 78 (4.5%) | 185 (10.6%) | 131 (7.5%) |
Source: Duke University Common Data Set filings, section B2, degree-seeking first-time first-year column. Percentages calculated against total first-year enrollment of 1,744, 1,742, 1,740, and 1,749 respectively. American Indian or Alaska Native and Native Hawaiian or Other Pacific Islander counts, each in single digits throughout, are omitted for space.
Read the unknown row before reading any other. Students declining to report race rose from 65 to 185 between Fall 2021 and Fall 2024 before settling at 131. That movement is large enough to absorb much of the change in the reported categories, and it makes precise year-over-year comparison of any single group unreliable. Institutions across the country saw the same effect after 2023, when race-conscious admission ended and reporting behavior changed alongside it.
With that caution in place, two movements are large enough to note. Asian enrollment rose from 18.9 percent to 25.1 percent of the first-year class across the period. Black enrollment fell from 143 students to 113, reaching 6.5 percent in the most recent filing, its lowest share in the window. Out-of-state share has held between 83 and 85 percent of the first-year class in every year Duke reported it.
Price, aid, and what students actually borrow
| 2021-22 | 2023-24 | 2024-25 | 2025-26 | |
| Costs reported for | 2022-23 | 2024-25 | 2025-26 | 2026-27 |
| Tuition | $60,435 | $66,325 | $70,265 | $73,740 |
| Required fees | $2,253 | $2,433 | $2,907 | $2,635 |
| Housing and food | $16,650 | $18,128 | $19,247 | $20,222 |
| Total published cost | $79,338 | $86,886 | $92,419 | $96,597 |
| First-years with need | 657 (37.7%) | 653 (37.5%) | 759 (43.6%) | 726 (41.5%) |
| Average need-based grant | $56,197 | $66,178 | $70,439 | $76,455 |
| Average total aid package | $62,270 | $71,878 | $76,028 | $83,513 |
| Share of need met | 100% | 100% | 100% | 100% |
Source: Duke University Common Data Set filings, sections G1, H2, and H2A. Each filing reports costs for the following academic year, so the column headings above give the filing year and the row beneath gives the year the prices apply to. Aid figures cover full-time first-time first-year students. Share of need met is Duke’s reported average for students awarded any need-based aid.
Published cost of attendance by component, Duke University
Published tuition rose 22 percent across the four filings, from $60,435 to $73,740, and total published cost rose a nearly identical 21.8 percent to $96,597. Over the same span the average need-based grant to a first-year aid recipient rose 36 percent, from $56,197 to $76,455. Aid outran price by a wide margin, and the most recent grant figure now exceeds published tuition.
Published tuition against average need-based grant for first-year aid recipients, Duke University
The share of the entering class with demonstrated need moved from 37.7 percent to 41.5 percent, with a peak of 43.6 percent for Fall 2024. Duke met 100 percent of demonstrated need in every filing. Those two facts together explain the borrowing figures better than the sticker price does.
Share of graduating class borrowing and average cumulative amount borrowed, Duke University
Borrowing has fallen while prices climbed. The share of graduates leaving with any loan debt dropped from 29.6 percent to 22.9 percent, and the average cumulative amount borrowed peaked at $27,788 in the 2023-24 filing before easing to $26,395. Fewer than one Duke graduate in four now borrows anything at all, and those who do carry less than the national average for a four-year private institution.
Class size and faculty
Duke reported 1,501 undergraduate class sections in Fall 2021 and 1,571 in Fall 2025, with a peak of 1,625 in Fall 2024. Sections enrolling fewer than 20 students accounted for 69.9 percent of the total in Fall 2021 and 71.0 percent in Fall 2025. Sections of 100 or more held almost perfectly steady at 45, 45, 45, and 42.
Full-time instructional faculty grew from 1,518 to 1,912 across the period, a gain of 26 percent against undergraduate enrollment that moved from 6,883 to 6,525. The reported student-to-faculty ratio reads 6 to 1 in 2021-22, then 5 to 1 in the next two filings, then 7 to 1 in 2025-26. That last figure reflects a change in the reporting definition rather than a change on campus: the 2025-26 form asks institutions to include graduate students and the faculty who teach them, and Duke’s calculation shifted from 6,490 students to 9,263. Comparing the 2025-26 ratio to earlier years measures a form revision, not a classroom.
What Duke students actually study
| Share of bachelor’s degrees | 2021-22 | 2023-24 | 2024-25 | 2025-26 |
| Computer and information sciences | 12.38% | 18.10% | 16% | 14.69% |
| Interdisciplinary studies | 2.27% | 1.60% | 11% | 14.18% |
| Social sciences | 16.98% | 15.60% | 11% | 13.00% |
| Engineering | 18.60% | 14.30% | 11% | 12.44% |
| Biological and life sciences | 15.31% | 12.60% | 13% | 10.13% |
| Public administration and social services | 10.53% | 6.00% | 6% | 6.33% |
| Health professions | 0.05% | 0.00% | 9% | 6.33% |
| Psychology | 5.94% | 5.50% | 5% | 5.30% |
| Mathematics and statistics | 4.68% | 7.00% | 6% | 5.19% |
Source: Duke University Common Data Set filings, section J1, degrees conferred in the year preceding each filing. Percentages use majors rather than headcount, so double majors appear twice. The 2024-25 filing reported whole percentages only. Categories follow federal CIP coding, and reassignment of a program between codes moves share between rows without any change in student behavior.
Share of bachelor’s degrees conferred by discipline, 2021-22 filing against 2025-26 filing
Computer science climbed from 12.38 percent of degrees to a peak of 18.10 percent and has since eased to 14.69 percent, and it is now the largest single category. Engineering moved the other way, from 18.60 percent to 12.44 percent, though the two categories together have held between 27 and 32 percent throughout. Mathematics and statistics rose then settled. Psychology barely moved.
Interdisciplinary studies jumped from 1.60 percent to 11 percent between the 2023-24 and 2024-25 filings and reached 14.18 percent in the most recent year, while health professions moved from zero to 9 percent over the same single interval. Movements of that speed do not happen because students change their minds. They happen because programs get recoded, and Duke has several majors that sit naturally in the interdisciplinary CIP family, neuroscience most prominently. Treat those two rows as reclassification rather than migration, and note that biological sciences fell by roughly the amount interdisciplinary studies gained.
Public policy, reported under public administration and social services, halved from 10.53 percent to around 6 percent and has stayed there. That decline is steady across three filings and is more likely to be real.
Beyond the numbers
The Common Data Set reports Duke as one institution. Applicants do not experience it that way. Duke admits separately to Trinity College of Arts and Sciences and to the Pratt School of Engineering, and the applicant chooses one on the application. Duke does not publish school-specific admit rates, so the effect is invisible in the CDS, but the two schools evaluate against different rubrics and Pratt engineering faculty participate directly in reviewing engineering applicants. Duke has publicly reported the split for individual rounds: for the Class of 2029, 1,552 regular decision admits went to Trinity and 401 to Pratt.
Movement between the two is asymmetric. Pratt students can transfer into Trinity through a straightforward internal process, while Trinity to Pratt is harder. An applicant who chooses Pratt is choosing the more selective door and keeping the easier exit. An applicant who chooses Trinity intending to switch into engineering later is making a bet that the university does not guarantee.
Duke also rates applicants on a numerical scale across six distinct criteria rather than the two-rating approach common at Ivy League peers, with the scale differing between Trinity and Pratt. Combined with the C7 movement described earlier, that structure rewards a coherent, program-specific case over a generically strong file.
On price, Duke made a commitment in June 2023 that the aid tables only partly capture. Undergraduates from North Carolina and South Carolina with family incomes of $150,000 or less receive full tuition grants, and those under $65,000 receive additional support for housing, food, and course materials without loans. Duke projected roughly 340 beneficiaries in the first year and additional spending of $6 million to $7 million annually over five years. The in-state application surge and admit-rate compression described above began in the cycle immediately following.
Two structural facts sit behind the whole record. Duke’s undergraduate enrollment has not grown, moving from 6,883 to 6,525 across the period even as faculty expanded 26 percent, so every gain in applications lands against a fixed number of seats. And Duke’s reporting has narrowed, with early decision counts, residency detail, waiting list figures, and one test range all absent from recent filings. Applicants relying on published data to calibrate expectations have less to work with each year.
What the five-year record means for applicants
The round matters more than the rate. Early decision admits accounted for 60 percent of the Fall 2024 entering class, and the approximate regular admit rate for that cycle was 4.2 percent against an early rate of 17.3 percent. An applicant who would attend Duke over any other option and whose family can commit without comparing aid offers is leaving a substantial advantage unused by waiting. An applicant who needs to compare packages should not apply early, and should build a list that does not depend on Duke.
Residency is a real and measurable factor. North Carolina applicants were admitted at 7.94 percent for Fall 2024 against 3.86 percent for international applicants. Carolinas families under $150,000 should treat Duke as a price-competitive option rather than a reach on cost, because full tuition grants change the calculation entirely.
Choose the school, not the odds. Trinity and Pratt admit separately, and an application that reads as engineering-shaped in the Trinity pool, or as undecided in the Pratt pool, works against itself. The internal transfer asymmetry favors applicants with a real engineering case who apply to Pratt.
Testing remains optional in practice and in weight. Roughly half of each enrolled class submitted an SAT score and Duke rates scores as merely considered. A student with a 1510 has a score worth sending. A student below the reported 25th percentile should weigh whether the score adds anything, because the transcript and the recommendations carry more of the decision than they did five years ago.
Finally, do not read the admit rate as a difficulty setting. Duke’s rate rose to 6.78 percent in 2023 because fewer people applied, not because the university relaxed. The class it built that year was the same size, with the same retention and the same academic profile, as every other class in this record. The number that actually changed was how many people were competing.
