As I post this, the embargo has been lifted, and articles about PISA 2025 are appearing on news websites around the world. And yes, there is plenty of bad news to report. In many education systems, the results are still moving in the wrong direction. A few years ago, it looked as if these results might mainly have been a Covid shock. However, it is becoming increasingly difficult to explain away the results as just that, PISA states. I do think that we need to remember that these fifteen-year-olds were 10 when COVID happened.

Over the past 24 hours, I have read so much that my head is just spinning. Apologies in advance to my students for the tiny eyes I will be bringing to class today. There is so much in PISA that I don’t really want to repeat the rankings and country stories you will find everywhere else today in the media. Instead, here is my own slightly idiosyncratic selection from the enormous grab bag of data and insights. Really, it’s a lot.

And, just to be clear: much of what follows is correlational. PISA can show us that things are related, but usually not that one thing caused another. Although, reading some of the accompanying texts, you can certainly sense that distraction by technology is being assigned quite a bit of the blame for some of the decline.

• Let’s start with what, for me, is the graph of PISA 2025:

  

  These are not the results of one particular country or education system. The graph shows the average performance of 23 OECD countries for which this longer trend can be followed. And the picture is hardly cheerful.

  This is also why education systems such as England, Ireland and Estonia will undoubtedly continue to attract attention. England and Ireland have remained remarkably stable, while Estonia has declined much less than many other high-performing systems such as Finland or Flanders. There are even a few countries where performance has actually improved. And by “improved,” I don’t mean moving up a ranking because everybody else got worse. Their students actually perform better than before. The catch is that most of these countries or regions still perform at relatively low levels.

  Still, perhaps the most important thing about this graph is the graph itself. We are getting further and further away from the idea that the decline we saw in PISA 2022 was simply a temporary COVID dip. The decline had already started before the pandemic in many places. And now, several years later, it continues.

• The OECD still invested quite some time of a lot of students and researchers in the concept of Growth Mindset. They started this last round, although the concept itself is, well, not that strong/relevant anymore. The OECD’s own interpretation has become considerably more nuanced in this round. Growth mindset is no longer presented as a stand-alone driver of achievement, but as one part of a much broader motivational profile.
• Then there is the environment. This was an important theme in PISA 2025, and the message is decidedly mixed. My attempt to summarise Andreas Schleicher’s summary would be this: climate change has become much more visible over the past decade, but young people say they know less about environmental issues.

  As with many items in PISA, this is self-report, which matters. But I do think the broader trend is striking. Compared with ten years ago, fewer students report being familiar with several important environmental challenges. This seems somewhat paradoxical. Over the past 10 years, climate change has been all over the place, yet that greater visibility has clearly not translated automatically into young people feeling more knowledgeable about concrete environmental issues.

  One striking example I saw in the data from the Netherlands. Dutch students report less enjoyment of science and a lower science self-concept than the average across the 14 EU countries used for comparison in the Dutch report. They are also less environmentally aware and participate less in environment-related activities. More remarkably still, Dutch 15-year-olds report the lowest sense of personal agency over environmental change of all countries participating in PISA.

• I also need to discuss this important methodological nuance in PISA. Schleicher drops this sentence into his paper almost casually:
  

  “Students across the OECD reported making less effort on the PISA test in 2025, compared to 2022 and 2018.”

  There are a couple of paragraphs about this in the main report too, including references to a table I couldn’t yet examine as closely as I would have liked (dead link because the report wasn’t published yet). I think this could matter a great deal. Do note – again- that effort is also self-reported, so one must be careful. However, when students from various countries report having put in less effort on the PISA test in 2025 than students in 2022 and 2018, it becomes a somewhat difficult methodological question. Are we just measuring the loss of knowledge and skills, or are we also measuring a reduced tendency to show off that knowledge and those skills in a low-stakes test? I am not sure whether this eliminates the decline at all, but it could make its interpretation tricky.

• There is another fascinating story hiding behind the averages: who exactly is going backward? For years, one of the most worrying trends in education has been the heightened vulnerability of disadvantaged students to declines in educational performance. It’s a trend that worsens inequality. PISA 2025 results now complicate that story. Across OECD countries with comparable data, the decline since 2018 has been particularly pronounced among the most socio-economically advantaged students. In reading, the socio-economic performance gap has actually become smaller on average.

  At first sight, that sounds like good news. Unfortunately, there is a rather important catch… This gap has not primarily narrowed because disadvantaged students suddenly started performing much better. It has narrowed because advantaged students have become worse. So it brings us to a rather uncomfortable question for discussions about educational equality. When should we be pleased that the achievement gap is narrowing?

  Equality through improving the outcomes of children from lower-SES families is obviously a good thing. Equality, because the more wealthy are falling towards the bottom, is something rather different. I haven’t checked all countries, but this pattern isn’t universal either. In some education systems, including Flanders and the Netherlands, socio-economic gaps have remained much more stable while overall performance has declined. Broadly speaking, different groups have been moving down together. I’m not entirely sure which of these stories should make us feel better.

And then there are the questions PISA 2025 has left me with. One thing I particularly want to explore over the coming days concerns England and the United Kingdom. In international comparisons, it is very tempting to talk about “the UK education system”. But, just as there really is no such thing as a single American or Belgian education system, there isn’t one British education system either. England, Scotland, Wales, and Northern Ireland have had quite different education policies for the past decade, to say the least. That makes the UK particularly interesting.

It looks like England does a little better than the United Kingdom overall. However, the difference does not seem very significant based on the limited data I have managed to find so far. The part that intrigues me most is the hint that England is both a top-performing country and one with relatively small differences among groups of students.

If that holds up when I dig further into the data over the next few days, it becomes a genuinely interesting case. Smaller inequality is one thing. Reducing inequality without bringing the strongest students down is quite another matter.

So much more to delve into… I’ll keep you posted

P.S. Apologies for the somewhat different style of this post, but this really was an enormous amount of reading and searching in record time. I wanted to have actually read the material myself. I am well aware that I have probably already forgotten things. A 377-page international report, an almost 80-page contribution by Schleicher, a 162-page Dutch report and a pile of Flemish slides turned out to be quite a lot to digest. 🙂