On 8 September 2026 the OECD published PISA 2025, its ninth round of testing 15-year-olds. More than 760,000 students in 91 countries and economies took part. The headline is blunt: across OECD countries, reading and mathematics scores are the lowest PISA has ever recorded. Reading has dropped 28 points since 2015 and mathematics 22, and the OECD counts roughly 20 points as one year of learning.
It is tempting to file this under 'teenagers and phones' and move on. But the most useful sentence in the whole report is about younger children: most of the drop in reading was already visible in primary-school tests years earlier. If you have a six-year-old learning to read, or teach a class of nine-year-olds, PISA 2025 is about you. Here is what the report says, what it does not say, and what parents and primary teachers can do this term.
What PISA 2025 found
Three numbers matter. Reading across the OECD fell 14 points since 2022 and 28 since 2015, about a year and a half of learning. Mathematics fell 9 and 22 points over the same periods. Science, the main subject this round, held steady since 2022. One in five 15-year-olds is now a low performer in all three subjects at once, up from 16% in 2022. A low performer in reading struggles to find the main idea in a text of moderate length; in mathematics, with the relationship between units and price on a shopping receipt.
The decline is wide: reading fell significantly in 56 of the 74 systems with comparable data. A few improved. The United Kingdom raised its science score since 2022, together with Costa Rica, the Slovak Republic and Türkiye, and Italy now sits above the OECD average in all three subjects. Singapore, Japan, Korea, Estonia and the Chinese jurisdictions remain far ahead. PISA 2025 also introduced a test of computational problem solving, part of a new 'Learning in the Digital World' module whose full results arrive in May 2027.
Why the report points at primary school
The OECD is careful about causes. School closures during the pandemic did not predict reading trends well, and performance was falling in many countries long before 2020. Instead it names a decline in focus and attention, less reading for pleasure, and more absence from school. Students did worst on tasks with longer texts that require careful reading and joining several pieces of information. 'Hasty reading', answering fast and wrong, nearly doubled to 9% of responses since 2018.
Then comes the sentence that matters for younger children. Comparing PISA with PIRLS, the international reading test for 9- and 10-year-olds, the OECD found that most falls in PISA reading were foreshadowed by falls in PIRLS at earlier ages. In its own words, the trend 'is already apparent in primary schools'. Reading, it adds, is the gateway through which most academic learning passes: the systems with the biggest reading drops also fell most in maths and science. Whatever a 15-year-old scores was largely decided at six, eight and ten.
What parents and primary teachers can do
Rebuild careful reading, early. Ten to fifteen minutes a day with a text slightly longer than comfortable, followed by one question that cannot be answered by skimming: 'why did she do that?' beats 'what colour was the hat?'. Read for pleasure, in whichever language is spoken at home. Keep phones out of the room while practising. PISA finds that students who report digital distraction in class score 11 points lower, and that leisure screen time at school beyond an hour a day goes with progressively lower scores.
Let children set small goals. Only 37% of 15-year-olds say they plan how they study, yet goal setters do better, and the OECD's advice is that schools should help students set their own goals rather than hand them down. In primary school that sounds like 'I want to read a chapter book by December' or 'I will get the seven times table right every day this week'. And treat maths as something that happens in the world: shopping, time, sharing, measuring. PISA's low performers fail exactly those everyday tasks.
Where Samio fits, and what it does not claim
Samio Learning is built for ages 4 to 12, not for 15-year-olds, so it does not prepare anyone for PISA and we will not pretend it does. What it does is work on the foundations the OECD says the problem grows from. Reading moves from phonics to short passages with literal, inferential and main-idea questions that adapt to the child's level. Maths practises number sense, fractions and word problems set in shopping, time, measurement and sharing. Science starts at eight and physics at nine, and the computing module builds the habit PISA's new test measures: breaking a problem into steps and checking whether the plan works.
Practice is built for attention rather than for screen time: no ads, no feed, nothing to scroll, and a parent-set daily time limit if you want one. Each child chooses their own daily goal and, on reaching it, answers one question from Sam: did your plan work? Parents see those answers, the minutes actually practised per subject, and a report that explains what a level actually means. Everything is available in seven languages, the same languages in which PISA 2025 made uncomfortable reading this month.
For schools: seeing the gaps before they become PISA statistics
A 15-year-old below Level 2 was, almost always, a nine-year-old whose gaps nobody measured. Samio for Schools maps every exercise a class does to your national curriculum, from the Dutch kerndoelen and the Common Core to the DfE programmes, the KMK standards and their Spanish, French, Italian and Portuguese equivalents, and shows teachers skills, competencies and assessment criteria by class and by student. Alerts flag who is falling behind before the report card does.
A short placement test shows where each child is actually working, and teachers can hand out curriculum-grounded challenges and lessons in minutes. Pricing follows the country's income level, so the schools in the systems with the steepest declines are not priced out of the response.
See how Samio for Schools shows curriculum gaps by class →
Frequently asked questions
Does Samio prepare children for the PISA test?
No. PISA tests 15-year-olds and Samio is a primary-school platform for ages 4 to 12. What Samio practises, careful reading, maths in everyday contexts, early science and computational thinking, is the foundation the OECD says PISA results are built on years earlier. There is no PISA-style test in the app.
What does 'low performer' mean in PISA?
A student below proficiency Level 2. In reading that means struggling to find the main idea in a text of moderate length; in mathematics, difficulty with a simple situation such as comparing two routes or converting a price; in science, trouble recognising the correct explanation for a familiar phenomenon. Across the OECD, 31% of students are below Level 2 in reading, 35% in mathematics and 26% in science.
Are screens and phones the cause of the decline?
The OECD says the data do not point to a single cause. It links the reading decline to a loss of focus and attention, less reading for pleasure, more absence and the shift to skimming on screens. Digital distraction in class goes with lower scores, while limited, purposeful device use for learning goes with higher ones. Ten structured minutes of practice is not the same thing as an open feed.
When does the next PISA data come out?
The Foreign Language Assessment is due in April 2027 and the full 'Learning in the Digital World' results, covering self-regulated learning and computational problem solving, in May 2027. The next full PISA cycle is 2029.
