A data-driven verdict · PISA · TIMSS · PIRLS · PIAAC
Is Polish education actually bad?
Poles complain about their schools. The data says the honest answer depends entirely on who and when you measure — and it’s more interesting than either “it’s a disaster” or “it’s fine”.
See the finding ↓
The finding
Poland’s standing collapses across the lifecourse
Put all three life stages on one scale: 0 is the average of 27 wealthy countries, and 1 is the typical distance between them. Poland runs from +0.63 at grade 4 (4th–5th of 24 countries) to −1.23 as adults, i.e. 25th of 27. That descent is the story.
The same shape, indicator by indicator
Every grade-4 assessment sits above the mean of these 27 countries; every adult one sits far below. Age 15 is the hinge: only +0.19 SD, 8th–9th of 27, so Poland’s teen advantage is modest once the comparison set is like-for-like.
The decisive test
It isn’t just the old communist-era cohorts
The obvious rescue for “our schools are fine” is that the weak adults were schooled under communism. Tested and refuted: the Poland–OECD gap is widest at 25–44, among adults schooled entirely by the modern system, and Poland’s 25–34s rank 30/31.
Excellence, honestly
Great at preventing failure, unremarkable at the very top
Does Poland cultivate the exceptional, or level everyone toward the middle? At age 15 the group lacking the basics is distinctly small, which is good news. The share of top performers, though, only matches the OECD average. That combination is the classic signature of levelling.
Real excellence, but niche
Compared with rich countries only, the lead of Poland’s best grade-4 pupils shrinks. Against the median country of the whole study their share is 2×, against the median rich country only about 1.3–1.5×. That is still above average, though not exceptional. Poland is world-class in one narrow field, competitive programming.
- IOI ~5th all-time by gold medals
- ICPC University of Warsaw, 2× World Champion
- IMO ~17th all-time (a notch below)
A plausible mechanism
Poland’s Karta Nauczyciela pay is flat, seniority-gated and nationally uniform: a 1%/yr tenure bump, no link to results, and among the OECD’s lowest absolute pay. Weak marginal reward for cultivating top pupils. A hypothesis, not proof.
Exceptional or typical?
Above-typical, but not a detached outlier
Rank every rich country by its age-15-to-adult gap. Poland is 3rd of 27, but the two above it (Singapore, South Korea) are there because their 15-year-olds are outstanding. Poland’s gap is adult-deficit-driven, and it’s the largest of the genuine catch-up cluster.
The same gap, on the map
Green where 15-year-olds outrank the country’s adults, crimson where adults outrank the young. Poland sits in a warm band with the Southern- and Eastern-European catch-up cluster; the Nordics run the other way, on strong adults, so a deep crimson.
A separate layer · outcomes
Elite talent, but not the machine to cash it in
Does the talent convert to high-end output? Mostly abroad. Domestic private unicorns come to one (DocPlanner), and the biggest Polish-founder wins, ElevenLabs and ICEYE, are foreign-domiciled. Poland punches below its weight.
Explanations · the school model
The Prussian model explains less than it is credited with
Poland’s school problems are usually blamed on the 19th-century Prussian school, said to suppress potential. Yet Poland does best at grade 4, exactly where the system is most “Prussian”. A government report itself states that every school system in the world descends from the Prussian one. If Singapore and Poland both have it, it cannot explain the difference between them.
- The three loudest charges do not hold up. Poland has one of the lowest statutory instruction times in the OECD. Summed over all years of schooling that is 5,304h against 7,642h, or about 663 against 849 a year, a 22% gap. 95% of teachers report freedom in choosing methods, the same as the EU average (94%). In PISA 2022 creative thinking Poland scores 34 points to the OECD’s 33, while Singapore, at 41, is clearly ahead.
- The real residue is in governance; teaching is free of it. 95% of teachers are free in how they teach. On curriculum flexibility they fall below the EU average (68% vs 76%), and on choice of assessment too (74% vs 82%). Only 9% feel they influence education policy, and 10% feel valued by society. Without freedom over curriculum and assessment, freedom of method is illusory, because the curriculum and the exam decide what pays off. Grades 5–8 only, self-reported.
- The one lever with causal evidence is the moment of selection. The 1999 reform delayed sorting by a year. Pupils who under the old system would have gone to basic vocational schools gained roughly 100 PISA points from it. The 2017 reform undid that mechanism. The claim concerns the calendar. It says nothing about the “spirit” of the school.
The entire alternative fits in a narrow strip
- Non-public schools enrol ~7.4% of pupils, and no analysis of their results controls for family wealth. They do better on the 8th-grade exam and clearly worse on the matura. Since the direction of the gap reverses, the results are measuring who the pupils are. They say nothing about the quality of the school itself. Non-public schools select and charge fees, while the local school must take every child in its catchment.
- Homeschooling covers 1.2–1.3% of pupils, its growth has already slowed, and its outcomes remain unmeasured. It is a story about a regulatory gap, visible in enrolment: 96.7% of homeschooled pupils are enrolled in non-public schools. The statistics office does not break it out, the exam board does not publish its results, and the education institute has not studied it. Even the Homeschooling Foundation concedes as much.
- Nobody has measured the Polish lesson in a comparable way. Poland took part in no international programme of recorded-lesson observation, neither the TIMSS Video Study nor the later Global Teaching InSights. There is no representative national observational study either. IBE and school inspection have observed lessons, but for other purposes. Poland also sat out TALIS 2018, so the comparable self-reported data comes from 2008, 2013 and 2024. The argument about the “Prussian lesson” therefore goes on without a measurement of the lesson itself.
- Six-year-olds in school are now a rounding error. The 2014–2016 reform made grade 1 compulsory for the whole cohort, and the act of December 2015 reversed it. In 2024/25 primary schools have 5,033 pupils aged six or under against 428,512 first-graders. That is at most one child in eighty-odd. It is an upper bound because the GUS category also covers younger children and pupils in every grade. The age-seven threshold is now set by practice as much as by the rule. Why so few parents enrol a six-year-old, the data does not say. Nor is it known what the reversal cost, because nobody has estimated it.
The real divide
Most pupils are on the vocational track
Most post-primary pupils are not in a general secondary school. Across schools for youth, excluding post-secondary, liceum holds 44.7% of pupils, technikum 40.5%, and vocational-entry and other schools 14.8%. Counting every route that ends in a vocational qualification, 54.5% are on the vocational track. Technikum carries most of that number and awards both the matura and a qualification, so “vocational” here does not mean “no matura”. Eurostat, restating the same national returns in ISCED categories, gives 54.4% against an EU average of 49.9%. That is not independent confirmation, because the Eurostat figure comes from GUS; its value is that every EU country is counted the same way.
- general secondary (liceum)44.7%
- technikum40.5%
- vocational-entry and other14.8%
- The same split returns in the results. In 2024 the matura pass rate was 92.5% in liceum, 85.2% in technikum and 28.0% in stage-II vocational schools. In 2025 it was 90.4%, 79.5% and 24.3% (CKE Sprawozdanie, counted after the resit). Read that last figure carefully: only 1,353 leavers of that track sat the exam at all, a self-selected minority. The press regularly falls into a trap here. The July preliminary numbers exclude the resit, so set against last year’s post-resit figure they show a collapse that did not happen. Papers are also not equated between years, because each is published in full and there are no anchor items. Even a correct comparison may therefore reflect the difficulty of the paper more than the skills of the candidates.
- The track decides how many foreign languages you get. Liceum runs at 1.99 languages per pupil, technikum at 2.00, and stage-I vocational schools at 1.02. The allocation is set by the framework-curriculum regulation. Liceum and technikum carry two modern languages, stage-I vocational schools one. Whether fewer language hours produce weaker command we do not know, because Poland has had no representative test of language competence since 2011.
One language or two, depending on the track
The honest verdict