# ESAT international candidates: what the official tables actually show

> Nearly half of ESAT candidates live outside the UK. UAT-UK's official tables give their scores by area of residence and by first language, module by module.

[Formatted version of this page](https://www.frontiervue.com/exams/esat/guides/international-candidates)

Close to half of all ESAT candidates live outside the UK, yet the question of how overseas candidates actually perform is usually answered by guesswork. UAT-UK's 2025/26 technical report prints the answer: candidates split by area of residence into UK, EU and Other, split again by first language into English and everything else, with means, standard deviations and four percentiles for every one of the five modules. This page sets out the key columns of both tables, adds the geography, test-centre and fee figures from the annual report, and repeats the report's own explanation of the first-language gap instead of pretending none was offered. The second half matters just as much. Other is not a country, a group mean is not a prediction for a person, and no score threshold can be read out of any of this.

## Key takeaways
- Overseas candidates are not a minority. Of the 14,193 candidates who took the compulsory Mathematics 1 module, 6,909 (49%) live in the UK, 777 (5%) in the EU and 6,507 (46%) outside both.
- The Other residence group's mean is at or above the UK group's on every module: 5.21 against 4.07 on Mathematics 1, 4.92 against 3.98 on Mathematics 2, 5.08 against 4.28 on Chemistry and 4.62 against 4.00 on Physics, with Biology's 4.54 against 4.43 called broadly comparable by the report.
- The first-language gap runs the opposite way to what most people expect. The report states it is smallest on Biology, where the non-English-first-language mean is 0.02 higher, and largest on Mathematics 1 at 0.96, again with the non-English group ahead (5.09 against 4.14).
- Benchmark against the right group. The whole-cohort Mathematics 1 median is 4.4 while the Other group's median is 5.0, and the 90th percentiles are 6.7 and 7.6. Reading an overseas score against the headline distribution flatters it.
- The report is not silent about the gap: it puts the non-English group's lead down primarily to a large number of candidates in China who are very strong at Maths and Physics, and its summary calls this a strong performance from candidates outside the UK. It still publishes no country-level scores, and Other still is not China.

### Quick answer: where overseas candidates sit on the ESAT

| Item | Value |
| --- | --- |
| Overseas share | UK candidates run 48% to 51% by module, Other 45% to 47%, and the EU 4% to 6% |
| Direction of the residence gap | On Biology the UK (4.43) and Other (4.54) groups are broadly comparable; on the other four modules Other is higher |
| First-language gap | Largest on Mathematics 1 at 0.96 (4.14 against 5.09), smallest on Biology where the non-English group is 0.02 higher; the report credits Chinese candidates |
| Is there a threshold | No. UAT-UK publishes no pass mark and the report contains no threshold of any kind |

> Every figure here comes from a printed table or sentence in UAT-UK's 2025/26 ESAT technical report and its 2025-26 annual report candidate data. Nothing is read off a chart image and nothing is derived across tables.

## Contents
1. Who the overseas cohort is, and how big
2. By area of residence, module by module
3. By first language: the size and the direction of the gap
4. Change the benchmark: who you are actually standing next to
5. What these numbers cannot tell you

## 1. Who the overseas cohort is, and how big

The ESAT is a global test, yet the size of its overseas cohort is rarely stated plainly. UAT-UK's technical report splits candidates by area of residence into UK, EU and Other, where Other covers everywhere outside the UK and the EU. On Mathematics 1, which every candidate takes, the 14,193 candidates break down as 6,909 in the UK, 777 in the EU and 6,507 elsewhere. The annual report adds that the tests ran in 129 countries.

- **46%** Share of Mathematics 1 candidates living outside the UK and EU (6,507 people)
- **49% / 5%** UK and EU shares on the same module (6,909 and 777 candidates)
- **8,745** Asia's line in the annual report's continent table, which covers UAT-UK's three tests and is not split by test
- **£130** Test fee outside the UK and Republic of Ireland, against £75 inside

- **Test footprint:** 129 countries (UAT-UK uses Pearson as delivery partner, with over 5,500 professional test centres across 180+ countries and territories)
- **Key markets outside the UK:** China, India, Singapore, Malaysia, Hong Kong (The annual report's list also includes the United States, the UAE, Türkiye, Spain and France, similar to last year)
- **ESAT bookings split near evenly:** October: 5,423 UK, 5,079 overseas (January runs 2,257 UK and 2,022 overseas. These are bookings, not delivered tests)
- **Fastest-growing markets:** Africa +42%, Asia +39% (Relative to the 2025 entry statistics, the annual report's largest increases were in African markets, closely followed by Asian markets)

## 2. By area of residence, module by module

|  | UK | EU | Outside the UK and EU (Other) |
| --- | --- | --- | --- |
| Mathematics 1 (compulsory) | Mean 4.07, median 4.0 | Mean 3.75, median 3.7 | Mean 5.21, median 5.0 |
| Biology | Mean 4.43, median 4.5 | Mean 3.60, median 3.6 | Mean 4.54, median 4.7 |
| Chemistry | Mean 4.28, median 4.2 | Mean 3.57, median 3.5 | Mean 5.08, median 4.9 |
| Physics | Mean 4.00, median 4.0 | Mean 3.56, median 3.4 | Mean 4.62, median 4.6 |
| Mathematics 2 | Mean 3.98, median 3.9 | Mean 3.68, median 3.6 | Mean 4.92, median 4.8 |
| Mathematics 1 group size and share | 6,909 candidates, 49% | 777 candidates, 5% | 6,507 candidates, 46% |
| Mathematics 1 observed score range | 1.0 to 9.0 | 1.0 to 9.0 | 1.0 to 9.0 |

The report gives its verdict in two sentences, both worth keeping. First, EU students are typically the weakest, UK candidates sit in the middle and candidates outside Europe are the strongest. Second, on Biology the UK group (mean 4.43) and the Other group (mean 4.54) are broadly comparable, while on the other modules Other candidates outperform UK candidates. The EU group is also the smallest sample of the three, at 134 candidates on Chemistry and 75 on Biology. This section of the report stops at that ranking. The section that does state a cause is the first-language one below.

> Other is not shorthand for Chinese candidates. The column covers everywhere outside the UK and the EU, and no country-level score data is published for anyone. Be precise about what the report does say, though: in its first-language section it names China, putting the non-English group's lead down primarily to a large number of candidates in China who are very strong at Maths and Physics. That is a stated cause, not a score, and this page will not convert it into one. The annual report lists China among the key markets outside the UK and counts Asia as the largest continent at 8,745 overseas candidates, a table not split by test, but market size converts into no country's mean either.

## 3. By first language: the size and the direction of the gap

|  | English as a first language | Other first language |
| --- | --- | --- |
| Mathematics 1 mean (N, share) | 4.14 (7,676 candidates, 54%) | 5.09 (6,517 candidates, 46%) |
| Biology mean | 4.43 | 4.45 |
| Chemistry mean (N, share) | 4.35 (1,774 candidates, 56%) | 4.93 (1,372 candidates, 44%) |
| Physics mean (N, share) | 4.10 (6,131 candidates, 56%) | 4.44 (5,159 candidates, 44%) |
| Mathematics 2 mean (N, share) | 4.06 (6,438 candidates, 53%) | 4.78 (5,675 candidates, 47%) |
| 90th percentile: Mathematics 1 / Mathematics 2 | 5.6 / 5.8 | 7.6 / 7.2 |

The report states both ends of this gap: it is smallest on Biology, where the mean for non-English-first-language candidates comes out 0.02 higher, and largest on Mathematics 1 at 0.96. It also states the cause. The result is primarily down to a large number of candidates in China who are very strong at Maths and Physics, which the executive summary generalises into a strong performance from candidates outside the UK, largest in the Mathematics modules. The report then adds an observation it calls especially interesting: even in subjects with a higher reading load, such as Biology, candidates whose first language was not English still performed better. On composition, the share with English as a first language runs 53% to 56% across the modules, against 65% for Mathematics 2 in the previous cycle.

### Item-level language bias: two flagged items across five modules

UAT-UK runs a differential item functioning (DIF) analysis on every form, with category C the most serious band. One Mathematics 1 item was flagged as showing DIF by first language, favouring candidates who speak English as a first language. Chemistry had two flagged items, one of which was flagged as preferring both English-first-language candidates and UK White candidates over UK non-White candidates. Mathematics 2 had no items showing significant category C DIF.

### Biology and Physics: flagged by gender, not by language

On Biology one item was flagged by gender, favouring candidates who identified as Man, and the report records no significant DIF by first language, UK school type or UK ethnicity. On Physics two items were flagged by gender, one favouring male candidates and one favouring female candidates, again with no significant DIF by first language, UK school type or UK ethnicity.

### The sample thresholds and bands behind the DIF analysis

UAT-UK's DIF analysis requires at least 50 responses per group and at least 200 in total; below 200 the sample is not large enough and DIF is not reported. The bands are A for DIF not significantly different from zero or with an absolute value below 1.0, B for significant DIF with an absolute value of at least 1.0 and below 1.5, and C for DIF significantly larger than 1.0 with an absolute value of at least 1.5. The two-item finding holds within those thresholds, not outside them.

## 4. Change the benchmark: who you are actually standing next to

|  | All candidates: 25th / 50th / 75th / 90th | Other group: 25th / 50th / 75th / 90th |
| --- | --- | --- |
| Mathematics 1 | 3.6 / 4.4 / 5.3 / 6.7 | 4.1 / 5.0 / 6.1 / 7.6 |
| Biology | 3.2 / 4.5 / 5.7 / 7.0 | 3.4 / 4.7 / 5.8 / 7.0 |
| Chemistry | 3.5 / 4.5 / 5.7 / 7.0 | 3.8 / 4.9 / 6.4 / 7.7 |
| Physics | 3.0 / 4.1 / 5.3 / 6.5 | 3.3 / 4.6 / 5.9 / 7.1 |
| Mathematics 2 | 3.3 / 4.2 / 5.4 / 6.5 | 3.7 / 4.8 / 6.0 / 7.2 |

**How an overseas applicant should use these tables**: Residence table, first-language table, whole-cohort table

- Step one: change the benchmark. The headline percentiles flatter an overseas score
  - Mathematics 1 median: 4.4 overall, 5.0 in Other
  - Mathematics 2 median: 4.2 overall, 4.8 in Other
- Step two: do not assume language is the bottleneck. Even the higher-reading-load modules did not reverse it
  - Mathematics 1: 5.09 for the non-English group, 4.14 for English
  - Two items across five modules were flagged as category C DIF by language
- Step three: pick modules by course requirement. Scores are not comparable across modules and there is no total
  - Biology drew 1,547 candidates for the cycle against 14,193 on Mathematics 1
  - The Other group is 709 candidates on Biology against 6,507 on Mathematics 1

Turn it into a plan. When you locate yourself, benchmark against the Other-group percentiles rather than the headline ones, above all on [Mathematics 1](/exams/esat/guides/mathematics-1) and [Mathematics 2](/exams/esat/guides/mathematics-2), where the two medians diverge most. For the whole-cohort distribution see [the score statistics page](/exams/esat/guides/score-statistics), for how the scale is built see [scoring](/exams/esat/guides/scoring), and for cohort size and geography see [candidate numbers](/exams/esat/guides/candidate-numbers).

## 5. What these numbers cannot tell you

### There is no pass mark and no score threshold

UAT-UK publishes no pass mark and the technical report contains no threshold. How each university weighs an ESAT score against predicted grades and the rest of an application sits outside the report's scope. Reading any percentile as a safe zone for overseas candidates is something readers add. The distribution is for locating yourself relative to this cohort.

### The Other group is not country-level data

Other covers everywhere outside the UK and the EU. No country-level score distribution is published anywhere. The annual report's geography table stops at continents: Asia 8,745, Europe 1,789, North America 650, Africa 173, Oceania 135 and South America 36, with no split by test. Its prose does name China, India and Singapore among the key markets, but that is market standing rather than scores. Where the technical report names China, it is stating a cause, again with no figure attached.

### Module scores cannot be added or averaged

The five modules are calibrated and scaled separately. The report states plainly that scaled scores are not directly comparable across modules and that there is no aggregate or total score. So the fact that overseas candidates are ahead on both Mathematics 1 and Physics cannot be fused into one overall figure. Universities always see the modules as separate numbers.

### Absolute scores do not carry across cycles

UAT-UK requires the scaling to be fixed to each cycle's October ability distribution, so the transformation is redone every cycle. The report states that this cycle's scaled scores cannot be directly compared with those from the 2024/25 cycle. Whenever you meet an overseas-group figure from another year, establish which cycle it belongs to first.

### A group mean is not a prediction, and the cause is quoted, not invented

In both tables the groups spread across most of the scale: on all five modules the Other group runs from 1.0 to 9.0, while the EU group tops out at 8.9 on Biology, Chemistry and Physics. A group average predicts nothing about any individual. On cause, the report is explicit about the first-language gap, a large number of candidates in China who are very strong at Maths and Physics, and this page repeats that. For the residence gap it offers only a ranking, so this page adds nothing. Neither should be read as a verdict on ability.

> One boundary to carry away: the report supplies score distributions, plus an explanation wherever it chose to give one. It supplies no admissions advice at all. Every figure on this page comes from a printed table or sentence in UAT-UK's 2025/26 ESAT technical report and its 2025-26 annual report candidate data; nothing is read off a chart image and nothing is derived across tables.

## Questions
**Does UAT-UK publish ESAT scores for Chinese candidates specifically?**

Not as scores. The technical report goes no finer than the three-way residence split (UK, EU, Other) and the two-way first-language split, and the annual report's geography table stops at continents (Asia 8,745, not split by test). Chinese candidates do appear in the report, though: it attributes the non-English group's 0.96 lead on Mathematics 1 primarily to a large number of candidates in China who are very strong at Maths and Physics. That is a sentence about cause with no mean or percentile attached. Any circulating average for Chinese candidates comes from neither document, and this page will not derive one.
**Does a non-English first language put you at a disadvantage on the ESAT?**

The official data runs the other way. On Mathematics 1 the non-English-first-language mean is 5.09 against 4.14, a gap the report puts at 0.96; Chemistry is 4.93 against 4.35, Physics 4.44 against 4.10 and Mathematics 2 4.78 against 4.06. Only Biology is level, with the non-English group 0.02 higher. The report explains the direction too: chiefly a large number of candidates in China who are very strong at Maths and Physics, and it points out that even Biology, with its higher reading load, did not run the other way. At item level, just two items across five modules were flagged as category C DIF by first language, one on Mathematics 1 and one on Chemistry, both favouring English-first-language candidates.
**What do overseas candidates pay, and where do they sit the test?**

The annual report gives the fee as £75 for candidates in the UK and Republic of Ireland and £130 for candidates sitting anywhere else. Delivery runs through Pearson, whose network covers over 5,500 professional test centres in 180+ countries and territories, and tests were taken in 129 different countries this cycle. On volume, ESAT bookings ran 5,423 UK against 5,079 overseas in October and 2,257 against 2,022 in January, close to even both times. Those are bookings rather than delivered tests.
**Does the Other group's higher mean mean overseas candidates are stronger?**

This page does not draw that conclusion. What the report says is that EU students are typically the weakest, UK candidates sit in the middle and candidates outside Europe are the strongest, and that on Biology the UK and Other groups are broadly comparable while on the other modules Other candidates outperform UK candidates. That describes scores. On cause, the report explains only the first-language gap, crediting a large number of candidates in China who are very strong at Maths and Physics, and says nothing about the residence gap; this page follows it. Other also merges dozens of countries and territories into one column, and its own scores run the full 1.0 to 9.0 on every module.
**Should I benchmark against the whole cohort or against the Other group?**

Look at both, but do not read only the headline column. The Mathematics 1 median is 4.4 overall against 5.0 in the Other group, with 90th percentiles of 6.7 and 7.6; Mathematics 2 runs 4.2 against 4.8 and 6.5 against 7.2. If you live outside the UK and the EU, using the whole-cohort percentiles alone will systematically flatter your position. Both columns give relative position only: UAT-UK publishes no pass mark and universities publish no fixed threshold.

## Official sources
- [UAT-UK ESAT Technical Report 2025-26, Table 8: scaled score summary statistics by area of residence](https://uat-wp.s3.eu-west-2.amazonaws.com/wp-content/uploads/2026/08/07161800/UAT-UK-ESAT-Technical-Report-2025-26.pdf)
- [Same report, Table 11: scaled score summary statistics by first language, with the prose on the size of the gap and its stated cause](https://uat-wp.s3.eu-west-2.amazonaws.com/wp-content/uploads/2026/08/07161800/UAT-UK-ESAT-Technical-Report-2025-26.pdf)
- [Same report, Table 5: whole-cohort and per-event distributions used here as the baseline, plus the scale-construction notes](https://uat-wp.s3.eu-west-2.amazonaws.com/wp-content/uploads/2026/08/07161800/UAT-UK-ESAT-Technical-Report-2025-26.pdf)
- [Same report, Tables 28 to 31 and the module-by-module DIF commentary: items flagged by first language, gender and ethnicity](https://uat-wp.s3.eu-west-2.amazonaws.com/wp-content/uploads/2026/08/07161800/UAT-UK-ESAT-Technical-Report-2025-26.pdf)
- [UAT-UK Annual Report 2025-26 Candidate Data, Tables 3 and 4: geography, test-centre network, fees and bursaries](https://uat-wp.s3.eu-west-2.amazonaws.com/wp-content/uploads/2026/08/07161742/Annual-Report-2025-26-Candidate-Data.pdf)
- [UAT-UK Annual Reports index (public entry point for all four 2025-26 reports)](https://esat-tmua.ac.uk/about-uat-uk/annual-reports/)
