Popinz Guide

PISA 2025: Four UK School Systems Compared With Ireland

28 September 2026 · Cross-sectional OECD PISA 2025 data and four UK national reports (associations, not causes); one meta-analysis of randomised trials, one quasi-experimental study and one cross-sectional study for research context · 7 min

DOI: 10.1787/73451bc5-en

The OECD published the first volume of PISA 2025 results on 8 September 2026. The United Kingdom averaged 511 points in science, 494 in reading and 488 in maths, against OECD averages of 482, 461 and 463. The OECD places the UK in the top ten in all three subjects, with science up 12 points on 2022 and reading and maths level.

That average blends four school systems run by four governments, each of which published its own national report. In science, England and Wales sit 40 points apart.

Check the sample before you compare

The OECD did not give the UK the asterisk it uses when sampling standards are missed, but its country note records that UK exclusions exceeded PISA’s tolerance and rose on 2022. It still judged all data fit for reporting. Excluded pupils, often those with special educational needs, sit no test. The rate reached 7.6% for the UK minus Scotland, against 4.9% in 2022 and a 5% ceiling, and 7.89% in Scotland, up from 6.57%.

Northern Ireland’s weighted school response rate, 83% after replacement, fell short of the 85% threshold, though its non-response analysis found a representative sample. In England, non-responding pupils were a little more likely to have been eligible for free school meals (FSM) or to receive SEN support. In Scotland, participants had slightly higher prior attainment than the full sample. If the missing pupils would have scored lower, the published averages sit a little high, though none of the reports puts a number on that effect.

Four systems, four sets of results

NationScienceReadingMathsMaths 2022
England516497492492
Northern Ireland492482473475
Scotland491488467471
Wales476459456466
OECD average482461463

England scored above the other three nations in all three subjects and was the only one whose science score changed significantly after 2022, rising from 503. Wales has the lowest scores of the four nations in every subject: 476, 459 and 456, against OECD averages of 482, 461 and 463. Its maths score fell for a second cycle running, from 487 in 2018 to 466 and then 456. Scotland’s maths average is lower than in any cycle from 2003 to 2018. In Northern Ireland, the maths gap between the highest and lowest-scoring tenth of pupils widened from 204 points in 2015 to 258.

Test dates and the GCSE year

PISA samples pupils aged 15 years 3 months to 16 years 2 months, so the point in the school year matters. England tested from February to May 2025, when about two-thirds of its sample were in Year 11 and sat GCSEs that summer. Wales and Northern Ireland tested in September and October 2025, with pupils due to sit GCSEs in 2026. Scotland tested in the same autumn window, earlier than in its previous cycles.

The OECD estimates the yearly learning gain around age 15 in England at 24.5 points in science and 27.0 in maths, with standard errors above 4 points, and warns against turning score gaps into months of schooling. English pupils were further into their GCSE courses, so the England-Wales gap is no clean measure of the two systems.

England: academies and free school meals

Academies taught 70.9% of the pupils sampled in England, maintained schools 16.7%, independent schools 7.7% and selective schools 4.7%. Volume 1 of the England report gives no scores by school type, so PISA 2025 offers no comparison of academies, multi-academy trusts and maintained schools; Volume 2 is due in December 2026. Eyles and Machin (2019) found that the sponsored conversions of low-performing schools in the 2000s raised end-of-secondary results for pupils already on roll. That is one quasi-experimental study of an early, selected group, and it cannot speak for the converter academies and trusts that followed.

In England’s sample, 24% of pupils had been eligible for FSM in 2025 or the previous six years. They averaged 474 in science against 522 for other pupils, 463 against 502 in reading and 448 against 497 in maths. On the OECD’s socio-economic index, England’s science gap between the most and least advantaged quarters (81 points) is close to the OECD average of 85.

Scotland: Curriculum for Excellence at 15

Scotland first implemented Curriculum for Excellence in 2011, according to the OECD’s 2021 review, with a Broad General Education to the end of S3 and a Senior Phase after it. Scottish scores held level with 2022, a stabilisation that the Scottish Government notes leaves results below earlier cycles. Background mattered less for reading: socio-economic status explained 3.6% of the variation in reading scores, lower than in any cycle from 2006 to 2022. The headteacher questionnaire missed the OECD participation standard, so treat Scottish staffing findings with caution. A single cross-sectional test cannot separate the effect of a curriculum from that of the pandemic or of changes in who sits the test.

Wales: the new curriculum has not reached PISA

The Curriculum for Wales began its phased roll-out in September 2022 and reached Year 10 in September 2025, the term in which Wales tested pupils already in Year 11. The Welsh report states that these learners were not exclusively taught using the new curriculum, so PISA 2025 tells you little about it.

Welsh socio-economic gaps are narrower than the OECD’s in every subject, at 65 points in science between the most and least advantaged quarters. Learners eligible for FSM in 2025 scored 439 in science against 481 for other learners. Low averages and narrow gaps coexist in Wales, and cross-sectional data cannot explain the combination.

Northern Ireland: grammar and non-grammar schools

Grammar schools accounted for 41.0% of the pupils sampled in Northern Ireland. Pupils eligible for FSM in 2025 scored 450 in science against 504 for other pupils, and 427 against 486 in maths, a 59-point gap. The reading gap between girls and boys widened from 14 points in 2015 to 30.

Ireland and Estonia as comparators

The England, Wales and Northern Ireland reports all compare with the Republic of Ireland, which scored 500 in science, 500 in reading and 480 in maths, with reading down 16 points on 2022. Irish testing ran in March and April 2025, close to England’s window, with 65.1% of pupils in Third Year and 27.9% in Transition Year. England’s reading score does not differ significantly from Ireland’s, while Northern Ireland’s sits significantly below it. Socio-economic status explains 9.6% of science variation in Ireland and 8.5% in the UK.

Estonia, a comparator for England and Northern Ireland, scored 527 in science and 508 in maths, both significantly above England. Its science gap between advantaged and disadvantaged students is 68 points, against 80 in the UK.

Phones, tutoring and families

In 2025, 80% of UK pupils attended schools that ban phones on the premises, up from 66% in 2022, against an OECD average of 49%. UK pupils in those schools were about 40% less likely to report digital distraction, and pupils who reported distraction in science lessons scored 38 points lower once the OECD accounted for socio-economic profile. Both are associations. The SMART Schools study (Goodyear et al., 2025) compared 1,227 pupils in 30 English secondary schools: restrictive policies went with 0.67 fewer hours of phone use during the school day, but with no difference in mental wellbeing or in total weekday and weekend use. It is a cross-sectional study, the weakest design cited here.

The OECD links study help such as peer-to-peer tutoring with higher science performance. Stronger evidence comes from a meta-analysis of randomised field experiments by Nickow, Oreopoulos and Quan (2024): tutoring improved learning by 0.29 standard deviations on average, with larger effects when teachers or paraprofessionals tutored and when sessions ran during school.

Family support fell. The share of UK pupils whose parents asked at least weekly what they did at school dropped from 81% to 78%, and the share with weekly talk about problems at school from 60% to 48%. For a school, the second figure is a reason to check that parents have a clear route for raising concerns with staff.

References