From the first PISA assessment in 2000 through the delayed 2022 cycle, the OECD ran eight assessment cycles covering reading, mathematics, and science. The major domain rotated across cycles, participation expanded, and the basis for long-run comparisons changed as each subject established its full reporting scale.
This historical series ends with PISA 2022. OECD published PISA 2025 Results (Volume I) on September 8, 2026, so 2022 is no longer the newest PISA cycle. PISA 2025 data are intentionally kept outside the table because the period covered here is 2000–2022.
Eight PISA Cycles From 2000 Through 2022
The cycle sequence was normally three years apart until the assessment planned for 2021 was postponed to 2022. Reading was the major domain in 2000, 2009, and 2018; mathematics in 2003, 2012, and 2022; and science in 2006 and 2015.
Cycle-by-Cycle Result Snapshot
Participation, the major assessment domain, the highest reported mean in that domain, and the published OECD benchmark for the same cycle.
Scroll horizontally to view all columns.
| PISA cycle | Major domain | Participation reported by OECD | Highest reported mean in major domain | Published OECD benchmark | Cycle note |
|---|---|---|---|---|---|
| 2000 | Reading | 32 in 2000; 11 more in the 2002 extension | Finland 546 | 500 | First PISA cycle; the reading scale was established. |
| 2003 | Mathematics | 41 | Hong Kong-China 550 | 500 | Mathematics became the major domain and its main trend scale was established. |
| 2006 | Science | 57 | Finland 563 | 500 | Science became the major domain and its main trend scale was established. |
| 2009 | Reading | 65 in 2009; 10 more in the 2010 extension | Shanghai-China 556 | 493 | Reading returned as the major domain; the extended administration is commonly reported as PISA 2009+. |
| 2012 | Mathematics | 65 | Shanghai-China 613 | 494 | Mathematics returned as the major domain. |
| 2015 | Science | 72 | Singapore 556 | 493 | Science returned as the major domain. |
| 2018 | Reading | 79 | B-S-J-Z (China) 555 | 487 | Reading returned as the major domain for the third time. |
| 2022 | Mathematics | 81 | Singapore 575 | 472 | Mathematics returned as the major domain; the assessment had been scheduled for 2021. |
The score column uses the highest published mean for the cycle’s major domain, not a claim of a precise statistical rank. Sampling uncertainty can make nearby country or economy estimates indistinguishable. Historical labels such as Hong Kong-China, Shanghai-China, and B-S-J-Z (China) follow the terminology used in the relevant OECD cycle reports and refer to the reporting geographies assessed in those cycles.
Participation Expanded, but Two Cycles Used Extended Administrations
Main-year participation rose from 32 countries and economies in 2000 to 81 in 2022. The first and fourth PISA assessments also added systems in follow-up administrations.
Countries and Economies in Each Main PISA Assessment Year
Main-year participant counts, 2000–2022
OECD also reports 11 additional systems in the 2002 extension to the first assessment and 10 additional systems in the 2010 extension to PISA 2009. The chart therefore shows main assessment years rather than total extended-cycle participation.
Participation growth changes the comparison universe over time. A country’s movement in a score-based ordering can therefore reflect both its own score and a different set of participating systems. Long-run performance comparisons are better read from matched trend series than from raw ordinal positions across cycles.
The 2022 Endpoint Broke Most Clearly From the Earlier Mathematics Pattern
OECD’s comparable 2018–2022 analysis uses a fixed set of 35 OECD countries for the headline average change.
- OECD mathematics performance stayed close to its 2003 level through 2018, then fell sharply in 2022.
- Reading and science had already turned downward before 2018 after reaching stronger OECD-average levels around 2012 and 2009, respectively.
- The 2022 assessment occurred after the cycle was postponed from 2021, but OECD trend analysis does not treat the pandemic as the sole explanation for the longer reading and science declines.
Why Every Subject Cannot Be Trended From 2000
A PISA score series becomes suitable for long-run comparison from the assessment in which that subject was first tested as the major domain under its full reporting scale.
The OECD benchmark in the cycle table is the published OECD average for that assessment’s major domain. For comparisons across time, OECD trend tables use consistent sets of member countries so that changes in OECD membership, participation, or missing observations do not create a false movement in the average. A cycle-specific benchmark should therefore not be chained mechanically into a long-run trend.
Data Sources and Measurement
All score, participation, cycle, and trend statements on this page are based on official OECD PISA publications or OECD PISA data documentation.
- OECD — Messages from PISA 2000
- OECD — Learning for Tomorrow’s World: First Results from PISA 2003
- OECD — PISA 2006: Science Competencies for Tomorrow’s World
- OECD — PISA 2009 Results, Volume I
- OECD — PISA 2012 Results, Volume I
- OECD — PISA 2015 Results in Focus
- OECD — PISA 2018 Results, Volume I: Mean Performance
- OECD — PISA 2022 Results, Volume I: How Countries Performed
Historical participant counts and the rotation of the major domains come from OECD’s PISA overview material. PISA 2022 included 81 countries and economies and about 690,000 assessed students.
OECD’s 2022 technical discussion explains why reading trends begin in 2000, mathematics in 2003, and science in 2006, and why consistent OECD-country sets are used for trend averages.
PISA 2025 is outside the date range of this historical page. OECD released Volume I of the PISA 2025 results on September 8, 2026.
Scores are reported as OECD-published mean estimates. Cycle-specific OECD averages, trend averages based on fixed comparison sets, and score-based ordering answer different questions and should not be treated as interchangeable measures.