# Getting into UT is only the first decision

## What admission really means, by residency, class rank and major, 2010–2026

Evidence cutoff: October 30, 2026. Entering years refer to fall cohorts unless a source says otherwise. This is a public-record reconstruction; missing rates remain missing.

## The number that answers the wrong question

UT Austin's published acceptance rate fell from **47.0% in 2010 to 22.2% in 2025**. During the same period, it made more offers and enrolled more freshmen. Applications grew much faster than either. The university became harder to enter by the usual aggregate measure, but that measure combines several different allocation systems. [Sources: UT Common Data Sets, Section C1](https://reports.utexas.edu/common-data-set/pdf).

A Texas student who meets the automatic-admission requirements holds a claim to a university place. That claim does not extend to the requested major. Another Texan competes for both university admission and a program. An applicant from outside Texas competes within a much smaller geographic allocation. After an offer, the student must still find that the program, cost and housing make enrollment possible. [UT review and decision process](https://admissions.utexas.edu/apply/review-decision-process/).

**The useful unit is a particular decision for a particular applicant group.** University admission, first-choice-major admission and enrollment are separate outcomes. The public data are unusually rich about the first and third, and unusually thin about the second.

This investigation joins sixteen Common Data Sets, the annual automatic-admission reports, archived college publications, an overlooked UT System admissions inquiry, state admissions files and fifteen years of high-school-to-college records. It produces a documented history rather than a calculator that invents personal probabilities.

Three findings change the usual account. Almost half the latest fall in the headline rate is attributable, arithmetically, to a change in the geographic composition of applicants. The share of Texas-high-school enrollees outside UT's published automatic-rank band has grown substantially since 2010, although the band itself changed. And the state's feeder-school records have become less complete in a way that can imitate a change in access.

## What a student can actually use

For fall 2025, UT's Common Data Set reports **35.9% in-state, 7.5% out-of-state U.S., and 7.9% international** university admission. The in-state figure includes automatic admits. It is not the rate for a Texan applying through holistic review. Nor is any of these a rate for Computer Science, business or engineering. [2025–26 CDS, C1](https://utexas.box.com/s/z5p8zzbr6yk4eluam3v3dt2wyaha3cja).

For eligible Texas students entering in **2026, 2027 or 2028**, the published automatic threshold is the **top 5%**. Eligibility includes the applicable schooling, curriculum, timing and application requirements; class rank in isolation is not the whole rule. A qualifying university guarantee still leaves the requested major under review. [UT eligibility guidance](https://admissions.utexas.edu/apply/review-decision-process/); [TEA's October 2026 notice for 2028 entry](https://tea.texas.gov/taa-letters/university-texas-austin-automatic-admission-policy-3).

The evidence tables in the companion data report the best matched public number for the selected decision. A published college-wide rate pools applicants across residency and rank. **Do not multiply it by a university-wide rate:** the two quantities are not independent probabilities, and major offers are part of the university's admission process. A group average also cannot measure the effect of one student's grades, school, essays or achievements.

## More applications, more places, a much lower rate

The published 2010 counts were **31,022 applications, 14,583 offers and 7,275 enrollees**. For 2025 they were **90,690, 20,154 and 9,900**. Applications rose **192.3%**; offers rose **38.2%**; enrollment rose **36.1%**. The admission rate fell 24.8 percentage points because demand as measured by applications outpaced the expansion of offers. [UT CDS archive](https://reports.utexas.edu/common-data-set/pdf).

That is strong evidence of rising aggregate competition. It is not evidence that UT simply held the entering class fixed or cut offers in half. Nor does an application count tell us how many students would enroll if offered their preferred major at an affordable price.

There is a material historical measurement problem. A 2014 UT System inquiry explicitly identifies the same early application totals as including **complete and incomplete applications**. The CDS convention calls for completed actionable applications. The series here preserves the university's published figures and flags the discrepancy; it does not silently rewrite the early years or assume later practice is perfectly comparable. [Admissions inquiry attachments, PDF pages 11 and 16](https://www.utsystem.edu/sites/default/files/documents/ut-system-reports/2014/ut-austin-admissions-inquiry/admissions-report-attachments.pdf).

For 2026, UT reports **98,841 applications and 9,487 new first-time freshmen**, but the announcement does not supply a corresponding offer count. Dividing enrollment by applications would produce a yield-adjusted enrollment ratio, not an acceptance rate. No 2026 admission rate is calculated here. Its twentieth-class-day enrollment snapshot also differs from some earlier census conventions. [UT's September 22, 2026 announcement](https://news.utexas.edu/2026/09/22/ut-sets-academic-performance-records-remains-the-no-1-public-university-in-texas/).

## Almost half the latest drop is a change in who applied

Between 2024 and 2025, the headline rate fell from **26.64% to 22.22%**, a decline of **4.42 percentage points**. Out-of-state U.S. applications rose from **23,015 to 35,268**, increasing their share of all applications from **31.6% to 38.9%**. This group has a much lower observed admit rate than the in-state group. [2024–25 CDS](https://utexas.app.box.com/s/d9izqb6s8dw2xxg5h5sunxyhrnef2ay6); [2025–26 CDS](https://utexas.box.com/s/z5p8zzbr6yk4eluam3v3dt2wyaha3cja).

I decomposed the change into the contribution of changing applicant shares and the contribution of changing admission rates within each published geographic category. A symmetric decomposition, which averages the two possible orders of changing those inputs, attributes **2.02 points, or 45.7% of the total decline, to applicant composition**. The remaining **2.40 points** come from lower within-category rates.

One counterfactual makes this concrete: with the **2024 applicant mix and the 2025 group admission rates**, the overall rate would have been **24.27%**, rather than 22.22%. This is an accounting exercise. It does not identify why more out-of-state students applied, and it does not imply that any group avoided increased competition. All three published groups had lower admission rates in 2025.

The estimate is conditional on treating the two years' published geographic categories as comparable. The 2024 CDS explicitly uses physical address at application; the 2025 table does not repeat that clarification. A verified coding change would require recomputing this result. These categories should not be treated as a perfect legal-residency panel.

The opposite of the usual story is therefore partly true: a university can report a much lower headline acceptance rate while making more offers and admitting a larger class. The composition of applications changes the headline even before the institution makes a different decision about an otherwise similar applicant.

## The cutoff is a forecasting device, not a fixed share of every class

The automatic-admission threshold moved from the top 10% for fall 2010 to the top 5% for fall 2026. It did not tighten every year. It loosened from 8% to 9% in 2012 and from 7% to 8% in 2016. The date UT announces a cutoff is also not its entering year: the move to 6% announced in 2017 applied to fall 2019. The dataset records the entering cohort. [UT automatic-admission reports](https://provost.utexas.edu/automatic-admission-reports/); [current TEA notice](https://tea.texas.gov/taa-letters/university-texas-austin-automatic-admission-policy-3).

SB 175, enacted in 2009 and first applicable to fall 2011, lets UT limit automatic admission by estimating in good faith how many rank-eligible offers will fill **75% of designated first-time Texas-resident enrollment capacity**. UT works through whole percentile bands and includes all eligible applicants in the final band. The statute also constrains nonresident offers by the number estimated to fill 10% of first-time enrollment capacity. Neither rule is a fixed percentage of all application files or all offers. [Texas Education Code §51.803(a-1), (j)](https://tcss.legis.texas.gov/resources/ED/htm/ED.51.htm#51.803); [SB 175 analysis](https://capitol.texas.gov/tlodocs/81R/analysis/html/SB00175F.htm).

The mechanism is a capacity forecast: expected enrollment is offers multiplied by expected yield. A smaller eligible rank band can be consistent with more high-school graduates, more eligible students choosing to apply, a change in their probability of enrolling, or a change in capacity. The cutoff alone cannot tell us which force dominated.

This distinction also explains why a realized enrollment share need not equal 75%. Forecast errors, different yields and the indivisibility of whole rank percentiles matter. Published outcomes should not be used to infer a violation of a forecast-based statute without UT's capacity designation and forecasts.

## The nonautomatic story runs in two directions

UT's annual reports show **989 enrollees outside the top-10% band in 2010**, out of 6,535 students from Texas high schools. In 2025 they show **2,941 outside the top-6% band**, out of 8,797. The outside-band share rose from **15.1% to 33.4%**. That contradicts a simple story in which automatic admission progressively consumes every available place. [UT SB 175 reports, Tables 1.5 and 1.6](https://provost.utexas.edu/automatic-admission-reports/).

But it does **not** establish that the same type of nonautomatic applicant had better odds. The boundary moved: a student ranked 8% counted inside the guarantee in 2010 and outside it in 2025. Applicant numbers for the two groups are not supplied. Even within the constant top-6% era, outside-band enrollment rose from 2,339 in 2019 to 2,941 in 2025; without applications, that remains a count of places taken, not an admission probability.

Yield is another reason offer shares and class shares diverge. In the 2025 Texas-high-school report, the top-6% group received **12,771 offers and enrolled 5,856 students**, a **45.9% yield**. The other group received **4,074 offers and enrolled 2,941**, a **72.2% yield**. The latter group occupied roughly one-third of the class in this reporting population after receiving roughly one-quarter of its offers.

The gap could reflect competing offers, major assignments, costs or differences in applicants. These aggregate counts cannot distinguish those explanations. In particular, low yield among rank-eligible students is not a count of people denied their requested major.

### An archived memo supplies real early rates

A 2014 admissions inquiry contains a much more useful cross-tab for 2009–2013: completed Texas-resident applications, automatic offers and nonautomatic offers, using the consistent “coded Texas Resident” definition (Method 1), rather than the memo's broader legal resident pool. For 2010, the calculation is **1,681 / (21,483 − 10,041) = 14.7%**. The corresponding rates are **22.2% in 2011, 23.7% in 2012, and 17.8% in 2013**. These are derived rates for completed applications, assuming automatically admitted applicants had complete files, consistent with the director's explanation that incomplete applications were not reviewed. The 2013 numerator excludes PACE offers. [Inquiry attachments, PDF pages 11, 15, 17, 18 and 25](https://www.utsystem.edu/sites/default/files/documents/ut-system-reports/2014/ut-austin-admissions-inquiry/admissions-report-attachments.pdf).

The inquiry's own all-received denominator produces lower numbers: 12.7%, 19.1%, 20.3% and 14.6%. For 2009–2013 pooled, removing incomplete files changes the derived rate from **15.79% to 18.62%**. This is a meaningful difference created by counting rules, before discussing student quality.

The same memo explicitly warns that **Texas high-school graduates and Texas residents are not congruent populations**. That rules out the tempting modern shortcut of subtracting SB 175 automatic counts from CDS in-state counts. The contemporary resident-by-automatic-status application denominator remains missing from the assembled public sources. Consequently, this research cannot honestly establish whether a nonautomatic Texan in 2025 had better or worse university-wide odds than an out-of-state applicant's 7.5%.

## Major admission is the missing center of the story

The public record does not support a complete table of precise current acceptance rates for every UT major. It does support a more useful distinction: matched college counts, approximate departmental statements, historical observations with definition breaks, and programs with no verified rate.

**McCombs publishes the clearest recent matched counts:** its flyer headed “2025 Admissions” lists **16,512 applied, 1,546 admitted and 1,010 enrolled**. The offer/application ratio is **9.36%**, consistent with the admission page's rounded 9%. Finance, Accounting, Marketing, Management and MIS do not have separate freshman entry rates in this process: students apply to the undergraduate business program and enter as unspecified business. Honors is a separate selection. [Official 2025 flyer](https://www.mccombs.utexas.edu/wp-content/uploads/2025/10/2025-MCombsBBA_Flyer-Digital-Sept25-1.pdf); [UT colleges and degrees](https://admissions.utexas.edu/explore/colleges-degrees/).

There is a caveat even here. McCombs' fall 2025 magazine separately reports **19,452 BBA applications**. No definition in the recovered material reconciles that with 16,512. It might reflect a different application population, but that explanation is unverified. The defensible statistic is the **9.36% ratio in the matched flyer table**; dividing its offers by the magazine's application count would silently join incompatible records. A 2024-uploaded flyer also has a cohort-label conflict with secondary accounts, so those data remain flagged. [McCombs magazine, fall 2025](https://news.mccombs.utexas.edu/wp-content/uploads/2025/11/MC_FALL-2025_FULLPDF_11.21.25.pdf).

For **Aerospace Engineering** and **Computational Engineering**, the department's current FAQ gives approximate acceptance rates of **6% and 10%**, respectively. It supplies neither a dated cohort nor the matched counts needed to reconstruct those rates. These are useful official planning estimates, not audited 2025 measurements. The **School of Architecture** says it admits about **120 of 1,000 freshman applicants**, approximately **12%**, again without a cohort date. This is a school-wide estimate, not a separate rate for architecture versus interior design. [Aerospace department FAQ](https://ae.utexas.edu/academics/undergraduate/undergraduate-faqs/); [Architecture freshman FAQ](https://soa.utexas.edu/undergraduate-admissions/freshman-faq).

**Plan II Honors** supplies another approximate rate: roughly **370 offers from 2,500 completed applications**, about **15%**. This is selection into an honors major, not the College of Liberal Arts as a whole. The FAQ does not date those counts to a single entering cohort. [Plan II admissions FAQ](https://liberalarts.utexas.edu/plan2/prospective-students/application/faqs.html).

Historical McCombs statistics and electrical/computer engineering impact reports add genuine observations. They do not create a clean annual panel for all engineering, and older engineering publications contain apparent scope or labeling changes that prevent a confident continuous series. The research package retains the observations and their warnings rather than connecting incompatible points into a smooth trend.

For **Computer Science**, a precise current freshman rate could not be verified in a matched public primary-source table. Repeated internet figures are not a substitute. The same applies to many other majors and colleges. “Not publicly verified in this research” does not mean easy, and does not prove that no unpublished UT table exists.

### A college can become much harder to enter without shrinking its class

Archived primary McCombs publications show **7,122 applications and 1,403 offers in 2012 (19.7%)**, **7,495 and 1,628 in 2013 (21.7%)**, and **6,503 and 1,651 in 2014/15 (25.4%)**. A mirrored 2010 advisory presentation gives 5,685 and 1,377 (24.2%); it remains separately labeled because the original hosted file was not recovered. The historical ratios were not a monotonic decline. [Archived 2012 UT page](https://web.archive.org/web/20130102214836id_/http://www.mccombs.utexas.edu:80/BBA/Prospective/Statistics.aspx); [archived 2013 UT page](https://web.archive.org/web/20131029032115id_/http://www.mccombs.utexas.edu:80/BBA/Prospective/Statistics).

The cleanest recent comparison is 2020 versus the 2025 flyer. Applications grew **8,885 to 16,512**, offers fell **1,695 to 1,546**, and enrollment barely changed, **995 to 1,010**. The reported admission ratio halved, from **19.1% to 9.4%**, while yield rose from **58.7% to 65.3%**. These are matched within-publication counts; cross-year comparability remains conditional on unchanged definitions, especially given the separate 2025 magazine discrepancy. [Archived UT infographic headed 2020](https://web.archive.org/web/20210502055411im_/https://www.mccombs.utexas.edu/~/media/Images/MSB/Program%20BBA/Infographics/Updated%20101119/InfoBBA2019AdmissionStats2x.png?h=600&w=1100); [2025 flyer](https://www.mccombs.utexas.edu/wp-content/uploads/2025/10/2025-MCombsBBA_Flyer-Digital-Sept25-1.pdf).

That illustrates two distinct pressures on admission: more applications per place, and fewer offers needed to fill each place when a greater share accepts. A stable enrolled class does not imply a stable admission rate.

### Capacity can move differently inside the university

McCombs' own strategic plan says UT evaluated and rejected a proposed undergraduate expansion in 2023. The proposal included growth in resident and nonresident enrollment and tuition revenue sharing. The plan then says McCombs reduced enrollment toward its 2019 baseline, approximately 950 students, for fall 2024. [McCombs strategic plan, completed initiatives](https://www.mccombs.utexas.edu/about/strategic-plan/).

This is direct evidence that university-wide growth and access to a particular college can diverge. It does not establish why the proposal was rejected, or isolate its causal effect on admission rates. But it gives the major-specific scarcity a documented institutional mechanism: the allocation of teaching capacity and revenue, not merely applicants competing against a single university standard.

The taxonomy also changes. UT approved a School of Computing within Natural Sciences for fall 2026, bringing together Computer Science, Statistics and Data Science, and Information. Its announcement is not evidence of a particular new admission rate. Longitudinal data need a program crosswalk, not just the school names in effect today. [UT announcement](https://news.utexas.edu/2026/02/19/ut-launches-new-school-of-computing-uniting-computer-and-data-science-statistics-information-disciplines/).

## What “admitted to UT, but not my major” means

An automatic university offer settles the campus decision, not the requested program. Depending on the case and current process, an applicant may be considered for a second choice or receive another program option. UT says it **may** consider an automatic-eligible applicant's second choice after denial from the first; this is not a promise that every applicant receives a full second-choice review. [UT admissions FAQ](https://admissions.utexas.edu/apply/frequently-asked-questions/).

For a student committed to a particular field, the relevant outcome is therefore **an offer for that field**, not simply an offer bearing UT's name. If a student enrolls in another major, switching later becomes a separate internal-transfer process with its own prerequisites, capacity and review. Prerequisite completion and a strong GPA may establish eligibility without guaranteeing a seat. The evidence table links the current program rules.

This is especially consequential for business, computing and engineering. It is also program-specific: the Statistics and Data Science department says internal transfer is expected to begin in fall 2027 and external transfer is unavailable. A generic “enroll first and switch later” plan can therefore fail before the student ever competes for a place. [Statistics and Data Science admissions](https://stat.utexas.edu/academics/undergraduate-major/admissions).

CAP must be kept separate from direct freshman admission to UT Austin. It begins at another participating UT System institution. Students who complete the requirements receive a transfer guarantee into Liberal Arts, with specific majors still subject to restrictions; they compete for other available majors. Architecture and Nursing are unavailable through CAP competitive transfer. Historical PACE offers also require separate treatment; the early inquiry explicitly excludes them from some counts. A pathway offer should not be added to ordinary campus offers without a clear definition. [UT CAP guidance](https://admissions.utexas.edu/apply/alternative-pathways-to-enrollment/cap/).

**How many students are admitted to UT but denied their first-choice major? The assembled public data do not answer it.** The required numerator is first-choice denials among university-admitted applicants, split by automatic status and residency. University offer totals, college enrollment totals and internal-transfer applications do not identify that quantity. This is the most consequential missing cell in the public history.

## The feeder-school dataset that should exist—and its trap

THECB publishes annual school-level records of high-school graduates' college destinations. I combined **2010–2024 into 21,880 school-year observations spanning 1,936 TEA school IDs**. Every school's reported destination counts reconcile to its graduate total. Stable IDs, rather than names alone, anchor the joins. [Official state series](https://www.txhighereddata.org/high-school-graduates/hsgradsenrolled/high-school-graduates-enrolled-in-higher-education-the-following-fall-by-high-school-county-school-district-high-school/).

The panel records enrollment matched by the state, not applications or admissions. Schools with 25 or fewer graduates are excluded. Destinations with fewer than five students are pooled, so an absent UT row means a suppressed or zero count, not a verified zero. The package preserves those cells as missing and supplies matched-record suppression bounds.

A second limitation is more damaging to trends. Some graduates cannot be tracked because their identifiers do not support linkage. Across included schools, this share rose from **4.42% in 2010 to 8.58% in 2024**. At Plano West it rose from **4.59% to 26.57%**; at Westwood in Round Rock ISD, from **5.02% to 18.43%**.

Plano West's separately reported UT enrollment fell from **82 to 58**. Its untrackable graduates rose from **41 to 334**. Those facts cannot establish that its actual UT pipeline shrank. Missingness need not explain the entire change to make the simple trend unreliable.

The largest documented 2024 UT counts include Westlake, Allen, Westwood, Plano West, Lake Travis and Vandegrift. They are useful starting points for investigation. Ranking them as the schools with the best admission “odds” would confuse school size, application behavior, student preparation, offers, enrollment choices and record linkage. Even a count per graduate remains an enrollment measure, not an admission rate.

The annual summary includes bounds for the top-20 and top-50 shares of **matched** UT enrollment. The intervals overlap across many years, and they do not cover untrackable or otherwise unmatched graduates, or excluded small schools. The data therefore support a searchable documented history more strongly than a definitive claim that UT's feeder system became more or less concentrated.

## A guarantee can exist without being usable

A new experiment supplies evidence on the part that enrollment tables cannot explain. Giani, Murphy, Flores and colleagues randomized 32 high-poverty, historically low-sending schools, covering 547 low-income students in the top tenth of their classes. Treatment letters proactively guaranteed existing tuition support, an $1,800 housing scholarship and campus housing placement; control letters encouraged application without the guarantees. For automatically eligible students, the average school-level fall 2024 enrollment share rose from **11% to 30%**. [Final article, published October 12, 2026](https://journals.sagepub.com/doi/10.3102/01623737261482127).

This selected, school-level experiment does not establish a statewide effect or separate the contribution of information from financial and housing certainty. It does show that a university guarantee can leave substantial access unused. The treatment's enrollment effect for students outside the automatic range was smaller and statistically insignificant.

Earlier research points to the same need to distinguish stages. Kapor's model of the 2002 UT/Texas A&M cohort attributes much of the top-percent policy's attendance effect to information; that historical model is not a forecast of today's major chances. Work on persistent feeder patterns and work finding benefits for newly included students are also compatible: a policy can help marginal students without replacing the schools that supply most of the class. [Kapor](https://www.nber.org/papers/w32372); [Cortes and Klasik](https://www.nber.org/papers/w28280); [Black, Denning and Rothstein](https://www.aeaweb.org/articles?id=10.1257/app.20200137).

The lesson from the Opportunity Insights work is methodological: separate who applies, who is admitted and who enrolls. Its findings about private Ivy-Plus colleges should not be imported as explanations for UT without evidence about UT's own process. [Chetty, Deming and Friedman](https://opportunityinsights.org/paper/collegeadmissions/).

## Where the existing coverage stops

The initial coverage audit found several different products often presented as if they answer the same question. CDS aggregators report a university-wide offer/application ratio. Residency explainers improve that number by showing geography, but typically leave automatic eligibility inside the Texas average. Admissions consultants discuss particular majors and publish useful historical material, but some current figures are explicitly estimates. Official college pages occasionally supply the best counts, often outside the university's central reporting system.

This research builds on those efforts. Tex Admissions had already published historical McCombs and Computer Science charts in 2017; reconstructing a college history is not an entirely new idea. The added value here is reconciling original observations, preserving cohort and denominator conflicts, joining them with rank and residence histories, and publishing the missingness rather than filling it with guesses. [Earlier McCombs analysis](https://www.texadmissions.com/blog/mccombs-business-statistics); [earlier CS analysis](https://www.texadmissions.com/blog/ut-austin-computer-science-admissions-statistics).

The current discrepancies matter. Two Tex Admissions pages estimate roughly 1,000 or 1,400 McCombs offers against about 16,500 applications, producing approximately 6% or 8%; the official matched flyer reports 1,546 and 16,512. That does not make every consultant observation wrong. It shows why an estimate, an enrollment count and an offer count must remain distinct. [Major recommendations](https://www.texadmissions.com/blog/2019/2/25/recommended-ut-austin-majors-rank-and-test-score-minimums); [business-specific article](https://www.texadmissions.com/blog/2025/6/18/applying-for-the-mccombs-school-of-business-and-canfield-business-honors-cbhp).

Some news accounts also describe the statutory 75% as a share of the entire freshman class. The legal denominator is designated first-time resident capacity and the rule uses an enrollment forecast. That small-sounding difference changes how every automatic/nonautomatic comparison should be read. [Example of the simplified account](https://www.texastribune.org/2024/09/16/ut-austin-top-five-percent-threshold/); [statute](https://tcss.legis.texas.gov/resources/ED/htm/ED.51.htm#51.803).

What is settled: university admission and requested-major admission differ; rank guarantees are cohort-specific; applications outgrew offers; current public major data are uneven. What remains contested or unresolved: the exact modern nonautomatic Texas rate, a comprehensive first-choice-major history, several official denominator conflicts, and the mechanisms behind differing enrollment yields. What the new joins add: a geography-standardized account of the latest rate decline, corrected early nonautomatic denominators, and quantified deterioration in feeder-record linkage.

## How the reconstruction can be checked

The research package retains original values, source URLs, report years, entering years, relevant tables/pages, formulas and caveats. Observations from different reporting systems stay in separate files. A missing application count is blank, never zero. Rate units are explicit in the column names and codebooks: CDS admission/yield fields are fractions; major-rate and feeder-per-100 fields are percentages. The 2022 CDS was recovered from a mirror of UT's document because the original Box link was inaccessible; that provenance is explicitly marked.

**University rate:** offers divided by applications in the same published C1 table. **Yield:** enrolled divided by offers. **Major rate:** program offers divided by matching program applications; no division across unrelated publications. **Feeder measure:** state-matched UT enrollees per high-school graduate, with separate suppression and trackability information. None is an individualized admission probability.

**The composition calculation:** write each year's headline rate as the sum, across published residency groups, of applicant share times group admission rate. The composition term uses the change in each share multiplied by the average of that group's two admission rates. The within-group term uses the change in its rate multiplied by its average share. These terms sum exactly to the overall rate change. There is no fitted model and no claim of causal identification.

**One table error and one unresolved cross-source discrepancy require separate treatment.** The 2025 automatic-admission report's Table 1.5b has visibly shifted regional labels/counts; the aggregate totals are independently corroborated by Table 1.6b, and the regional rows are excluded. The current THECB admissions workbook also differs substantially from the CDS: for 2025 it gives 77,639 total applications versus 90,690 in C1. Its reporting manual describes complete, degree-seeking summer/fall applications and legal residence. That does not fully reconcile the gap. The workbook is retained as a separate series, not used to patch the CDS. [THECB admissions series](https://www.txhighereddata.org/enrollment/applicants-acceptance-and-enrollment/).

Confidence is strongest in arithmetic reproduced from matched tables and in statutory thresholds documented by entering cohort. It is lower for undated departmental approximations and any inference requiring stable reporting definitions. The main findings would change if UT supplies corrected tables, demonstrates a residency coding break, or releases a matched application-by-major cross-tab. The published calculations are structured to make those revisions straightforward.

### The records that would complete the history

The missing dataset is modest in concept: for each entering cohort, residency category, actual automatic-admission status and first-choice major, report applications, complete applications, university offers, first-choice offers, other-major offers, pathway offers and census enrollment. Add second-choice major, stable high-school and program IDs, rank bands, and historical crosswalks. Distinguish eligibility from actual admission authority, and record requested versus assigned major separately.

UT should also provide the capacity and yield forecasts used to set each automatic threshold. Those would distinguish policy pressure from demographic growth, application changes and shifts in enrollment behavior.

No individual student records are needed to answer the central questions. The public download contains the evidence tables and methods.

The appropriate advice to a student is therefore conditional. Identify the guarantee, if any; identify the actual program entry point; use only a rate with a matching cohort and decision; and evaluate any alternative-major or transfer route as its own competitive process. The public record can narrow uncertainty substantially. It cannot yet supply the exact requested-major odds that most acceptance-rate pages imply they know.