Yesterday, I wrote about the new PISA 2025 results, and my post ended with a question to explore: what happens if, instead of looking at the United Kingdom (UK) as a whole, as reported in the OECD’s first Volume, we look at England, Scotland, Wales, and Northern Ireland separately? So, this post delivers on that promise.

First question to answer: why is the UK so relevant? Just as there is no such thing as Belgian or even American education, there is no such thing as British education either. England, Scotland, Wales, and Northern Ireland have been pursuing their own education policies for quite some time. In recent years, this has almost become a natural experiment, as their choices have diverged quite substantially on some points. And then the second question becomes whether we can see any of this reflected in their PISA results.

This graph from the Financial Times summed it up rather provocatively yesterday: “England has bucked the international trend of collapsing school test scores.”

2) England is a rare exception.During the 2010s many other countries (including the other 3 UK nations) changed tack, experimenting with ideas like transferrable skills instead of focusing on tried-and-tested disciplines and methods.It looks like those experiments have failed.

— John Burn-Murdoch (@jburnmurdoch.ft.com) 2026-09-08T10:56:25.172Z

But… that graph may be putting it a little too strongly. I will come back to that later. But it’s clear something is going on.

In PISA 2025, England scored 516 in science, 497 in reading, and 492 in mathematics. By contrast, Scotland scored 491, 488, and 467, respectively. Northern Ireland scored 492, 482, and 473. Wales has the lowest scores of the four in each domain, with 476, 459, and 456. England’s scores are clearly higher across all three domains than those of the other three parts of the United Kingdom. This is a difference I couldn’t trace in the overall UK figures I had on Monday.

These differences are particularly large in mathematics and science. England is about 25 points ahead of Scotland in mathematics and science. If you compare this with Wales, the gap rises to 40 points in science, 38 in reading, and 36 in mathematics. A rough rule of thumb is that 20 points correspond to about a year of schooling. But I need to add that the OECD is the first to warn against interpreting score differences this way.

And then there are the different trajectories in the aforementioned graph. Between 2015 and 2025, England remained statistically stable across all three traditional PISA domains, while the OECD average has, in fact, been declining for some time. In mathematics, Scotland, Wales, and Northern Ireland all declined significantly over the same period. In reading, Wales and Northern Ireland declined significantly, while England and Scotland remained statistically stable.

Bear in mind that stable does not mean that nothing happened during the past 25 years. England’s mathematics score of 492, for example, is significantly below the 504 recorded in 2018. But there is no significant difference compared with 2015, and between 2022 and 2025, the score remained exactly 492.

But as I wrote in the introduction, the four regions have made different choices over the past 20 years. England, particularly since the reforms introduced from around 2010 onwards, has placed much greater emphasis on subject content, knowledge, a clearer national curriculum, systematic phonics, external assessment, and fairly explicit expectations about what pupils should know. For the Flemish readers of this blog, much of this will sound rather familiar.

Scotland, by contrast, opted with Curriculum for Excellence for a much stronger emphasis on broad competences, interdisciplinarity, and flexibility. Wales moved in a similar direction with Curriculum for Wales, towards a more purpose-driven curriculum, with broad goals and considerable room for schools to shape the curriculum themselves.

And Northern Ireland? I was there myself only two weeks ago, and I noticed that something is happening. Call it a remarkable change of direction. The proposed curriculum for 2028 literally describes itself as “knowledge-rich, coherent and ambitious”. It is intended to specify much more clearly what children learn and when, with content built up cumulatively and carefully. The reform is being led by Christine Counsell, who also contributed to the new minimum attainment targets in Flanders, with Lucy Crehan, an old acquaintance of mine, as vice-chair of the curriculum taskforce.

The conclusion now seems easy to draw. England placed greater emphasis on knowledge, clearly specified subject content, and remained relatively stable. Wales declined in reading and mathematics, Scotland in mathematics. Northern Ireland also lost ground and is now explicitly moving towards a more knowledge-rich curriculum. Case closed? You don’t know me very well if you think so. Things are usually… more nuanced.

Remember, we are talking about PISA. These are regions, not randomised control groups. I can come up with hundreds of things that potentially differ between Wales, England, and Scotland besides their curricula. What about demographics, economics, teacher shortages, assessment policies, school structures, absenteeism, and so on. It is impossible to conclude from these figures that curriculum policy caused these different trajectories. I do think personally that it may certainly have played a role, but PISA data do not allow causal conclusions, remember?

And reading the four regional reports, combined with some additional digging here and there, taught me quite a bit. The Welsh Government itself stressed that the pupils in PISA 2025 are the final PISA cohort who had not yet been taught under the new Curriculum for Wales. So it was the old curriculum that was being ‘tested’. It is therefore difficult to claim that Wales’s poor PISA results prove that Curriculum for Wales has failed. There is a small chronological problem with that argument. At the same time, Wales is showing a slow but steady movement towards greater emphasis on clarity, coherence, sequencing and the re-use of prior knowledge.

Story finished? No. Yesterday, I also wrote that the preliminary information I had received suggested that England was performing well with less influence from social background. Is that actually true, and what about the other regions?

There is still a substantial SES gap in England. The gap is 81 points in science and 89 in mathematics, roughly comparable with the OECD average. But interestingly, in reading the English gap of 66 points is significantly smaller than the OECD average of 78. At the same time, England has quite a lot of top performers. In mathematics, 14% reach Level 5 or 6, compared with an OECD average of 8%.

When we look at Wales, we see relatively small socioeconomic differences. Between the highest and lowest SES quartiles, the gap is 65 points in science, 50 in reading, and 70 in mathematics. All three gaps are significantly smaller than the OECD average. SES also explains relatively little of the variation in performance in Wales. But we should not forget that Wales combines this with low average performance.

Scotland makes the equity story even more complicated. The relationship between SES and reading performance became remarkably small in 2025. SES explains just 3.6% of the variation in reading performance, and the SES gradient is 19 points. Both are significantly lower than in previous PISA cycles.

So no, we cannot simply say that England chose excellence while Wales and Scotland chose equality. The results are much more complex (and interesting) than that.

While I was trying to make sense of all this, someone sent me an observation about the English sample. One of the advantages of sharing so much through this blog is having a very attentive and engaged group of readers. Thank you for that! The English report contains an extensive analysis of the pupils who did and did not participate in PISA. And it turns out that the non-response certainly doesn’t look random.

Among participating pupils who could be linked to England’s National Pupil Database, 1.4% had an Education, Health and Care Plan. Among selected pupils who did not participate, the figure was 7.1%. For other SEN support, the figures are 10.7% versus 25.0%. And 23.8% of participating pupils were eligible for free school meals, compared with 39.6% of non-participants.

What does this mean in concrete terms? Among the pupils who did not participate, there are relatively many pupils whom we might reasonably expect, on average, to have achieved lower scores. The English PISA researchers acknowledge the problem themselves, as they should. They conclude that the participating schools are reasonably representative, but identify potential bias related to individual pupil characteristics, including free school meals and SEN.

And there is more going on with the data. The researchers could not investigate whether participation was related to prior attainment because this cohort did not sit the relevant Key Stage 2 tests due to Covid. On top of that, 410 participating pupils could not be linked to the National Pupil Database and are therefore missing from this analysis.

Mind you, the OECD did judge the UK data suitable for reporting, although the UK exceeded the permitted threshold for exclusions. England itself met the formal PISA response standards, with a 90% school response rate and an 83% pupil response rate. Nevertheless, the English report explicitly warns that international and historical comparisons should be interpreted with caution. Which is why it is always a good idea to read everything.

I was also told that this issue could account for a difference of up to 10 points. I have not been able to confirm this so far, so I am not treating that figure as an established correction. But even if a bias of roughly that magnitude existed, it would not explain the full gap between England and Scotland, let alone Wales. It would, however, make the English story somewhat less spectacular.

Which brings me, finally, to the question several journalists have asked me over the past 24 hours: should England now be the education system for everyone else to follow? Well, as a system that has held up relatively well during a period of widespread international decline, England certainly offers an interesting and potentially instructive story. But we should not forget what happened to previous educational role models. Remember Finland?

Nor should we assume that curriculum explains everything. I happened to be in London myself on Saturday, and the big topic there is no longer simply knowledge-rich education, but… inclusion.