PISA 2025 Denmark: what the country note tells educators
Denmark’s PISA 2025 scores fell in maths, reading and science. We unpack the OECD country note: proficiency, equity, AI chatbot use, phone bans and more.
Key takeaways
- Denmark scored 471 in mathematics (above the OECD average of 463), 460 in reading and 478 in science (both close to average), and 506 in computational problem solving.
- Scores fell in all three subjects after 2022 and were lower than in any assessment before 2022.
- Since 2015, the share of low performers rose by 15 percentage points in mathematics, 14 in reading and 9 in science.
- Denmark has few staff shortages and high device use, which suggests access alone does not explain outcomes.
- 48% of students use AI chatbots to help them learn at least weekly, and 65% attend schools that ban phones, up from 40% in 2022.
Denmark has taken part in PISA since 2000. Its PISA 2025 country note, written by Francesco Avvisati of the OECD Directorate for Education and Skills, gives a detailed picture of how Danish 15-year-olds perform and learn. In 2025, 5 098 students in 283 schools completed the assessment, representing about 56 500 15-year-olds. The OECD notes that Denmark’s school response rate and student exclusions did not fully meet PISA standards, but after further analysis all data were considered fit for reporting.
Scores and trends
| Domain | Denmark 2025 | OECD average 2025 | Denmark change since 2022 | OECD change since 2022 |
|---|---|---|---|---|
| Science | 478 | 482 | −16 | −3 |
| Reading | 460 | 461 | −29 | −14 |
| Mathematics | 471 | 463 | −19 | −9 |
| Computational problem solving | 506 | 500 | New in 2025 | New in 2025 |
Danish students scored above the OECD average in mathematics and computational problem solving, and close to the average in reading and science. However, average performance was lower in 2025 than in 2022 in all three core subjects, and lower than in any assessment before 2022. From 2022 to 2025, the gap between the highest-scoring 10% and the lowest-scoring 10% widened, particularly in mathematics.
A strong floor, fewer students at the top
Denmark’s profile combines a relatively large share of students reaching baseline proficiency with fewer top performers than the OECD average.
| Domain | Level 2 or above (Denmark) | Level 2 or above (OECD) | Level 5 or 6 (Denmark) | Level 5 or 6 (OECD) |
|---|---|---|---|---|
| Science | 75% | 74% | 5% | 7% |
| Mathematics | 72% | 65% | 5% | 8% |
| Reading | 71% | 69% | 4% | 6% |
The long-term trend is the bigger concern. Compared with 2015, the proportion of low-performing students (below Level 2) increased by 15 percentage points in mathematics, 14 in reading and 9 in science.
Equity
Socio-economic background matters less in Denmark than on average. Advantaged students outperformed disadvantaged students by 69 points in science, compared with 85 points across the OECD. Socio-economic status explained 10% of the variation in science performance (OECD average: 12%), and 13% of disadvantaged students were academically resilient (OECD average: 12%). Between 2015 and 2025, performance declined to a similar extent among advantaged and disadvantaged students.
Curiosity, effort and learning habits
The attitudes data may be the most useful part of the note for teachers. In Denmark, 66% of students said they are curious about many things (OECD: 73%), down 8.7 percentage points since 2022. Half said they apply additional effort when work becomes challenging (OECD: 60%), and 22% agreed that school has been a waste of time (OECD: 24%).
These attitudes are strongly linked to performance in Denmark. After accounting for socio-economic profile, a one-unit increase in the curiosity index is associated with 30 more score points in science (OECD: 19), and the perseverance index with 25 points (OECD: 13). Some 68% of Danish students have a growth mindset (OECD: 69%).
Self-regulation has room to grow. Only about 33% of Danish students say they often plan their approach to studying, and 37% ask themselves questions to check their understanding (OECD averages: 37% and 49%). Help-seeking is a strength: 84% ask teachers for help and 83% ask classmates (OECD: 77% and 82%), and help-seeking is associated with 17 more science points in Denmark, compared with 6 on average.
Devices, phones and AI chatbots
Danish students report spending 3.2 hours a day using digital devices at school for learning (OECD: 1.7 hours) and 0.9 hours for leisure (OECD: 1.1 hours). Some 36% said classmates are often distracted by devices in science lessons (OECD: 28%), although Denmark is among the minority of systems where this distraction is not associated with a significant difference in performance.
Phone policies have changed quickly. In 2025, 65% of Danish students attended schools where phones are not allowed on the premises (OECD: 49%), up from 40% in 2022.
- 48%
- use AI chatbots to help them learn at least weekly (OECD: 46%)
- 37%
- use AI for preliminary research on a new topic (OECD: 31%)
- 31%
- use AI to summarise a text they had to read (OECD: 30%)
- 26%
- use AI to draft text for writing assignments (OECD: 29%)
School life
- Staffing: only 5% of students were in schools where principals said a lack of teaching staff hindered instruction (OECD: 39%), down from 10% in 2022.
- Teacher support: 66% of students said the teacher shows interest in every student’s learning, 66% that the teacher continues until students understand and 68% that the teacher gives extra help when needed. In 2015 the shares were 70%, 74% and 74%.
- Bullying: 25% of students reported at least one form of bullying a few times a month or more (OECD: 20%), an increase since 2022.
- Family support: 75% said parents or guardians ask weekly what they did at school, and 44% have weekly discussions about problems at school, down from 80% and 58% in 2022.
- Attendance: 28% skipped at least one day in the two weeks before the test (OECD: 22%) and 60% arrived late (OECD: 50%). In 2022 the shares were 20% and 49%.
Lessons for educators beyond Denmark
Denmark shows that good staffing and plentiful devices do not guarantee rising achievement. What happens during lessons, and how students approach learning, matters. Five ideas stand out:
- Make curiosity visible. Open units with a genuine question and give students room to wonder before explaining.
- Reward productive struggle. Plan tasks at the right level of challenge and praise strategies, not only answers.
- Teach self-regulation explicitly. Model how to plan a study session and how to check understanding with self-questions.
- Signal support. Small moves, such as checking in with every student and re-explaining when needed, are what students report on.
- Guide AI use. Nearly half of students already use chatbots every week. Teach them to evaluate what AI produces with activities like our AI literacy classroom activities.
Frequently asked questions
How did Denmark perform in PISA 2025?
Denmark scored 478 in science, 460 in reading, 471 in mathematics and 506 in computational problem solving. This was above the OECD average in mathematics and computational problem solving and close to the average in reading and science.
Did Denmark’s PISA scores go down?
Yes. Average results in 2025 were lower than in 2022 in mathematics, reading and science, and lower than in any PISA assessment before 2022.
How many Danish students use AI chatbots?
In PISA 2025, 48% of students in Denmark reported using AI chatbots to help them learn at least once a week, compared with an OECD average of 46%.
Do Danish schools ban mobile phones?
In 2025, 65% of Danish students attended schools where phones were not allowed on school premises, up from 40% in 2022. The OECD average was 49%.
Sources
- OECD (2026), PISA 2025 Results (Volume I): Denmark country note ↗
- OECD (2026), PISA 2025 Results (Volume I): Executive summary ↗
Figures are adapted from OECD (2026), PISA 2025 Results (Volume I): Future-Ready Students, OECD Publishing, Paris, https://doi.org/10.1787/73451bc5-en, published under CC BY 4.0. This is an adaptation of an original work by the OECD. The opinions expressed and arguments employed in this adaptation should not be reported as representing the official views of the OECD or of its Member countries. SparkSchool AI is not affiliated with or endorsed by the OECD.