The United States recorded mean scores of 502 in science, 463 in mathematics, and 490 in reading in PISA 2025. The latest cycle extends a U.S. series that begins in 2000 for reading, 2003 for mathematics, and 2006 for science.
United States PISA 2025 Scores
Science was the major domain in PISA 2025. Mathematics and reading were minor domains, and all three core-domain results are reported on their established PISA scales.
2025 scores, OECD comparison, and change from 2022
The change column uses the OECD’s published 2022-to-2025 change estimate rather than a subtraction of rounded display scores.
| Domain | U.S. 2025 | OECD average 2025 | U.S. − OECD | Change from 2022 | OECD trend assessment |
|---|---|---|---|---|---|
| Science | 502 | 482 | +20 | +2 | About the same as 2022 |
| Mathematics | 463 | 463 | 0 | −2 | About the same as 2022 |
| Reading | 490 | 461 | +29 | −14 | Lower than 2022 |
OECD comparison statements account for statistical uncertainty. A zero or small arithmetic gap should not be treated as a separate statistical finding.
- Reading moved differently from the other two core domains between 2022 and 2025: the OECD reports a 14-point decline, while mathematics and science were about the same as in 2022.
- The 2025 mathematics mean equals the OECD average at 463. Reading and science are above their corresponding OECD averages.
U.S. PISA Scores Across Every Comparable Cycle
PISA is cycle-based rather than annual. The comparable trend starts when each domain first served as a major PISA domain: reading in 2000, mathematics in 2003, and science in 2006.
Reading, mathematics, and science trends, 2000–2025
Published U.S. mean scores by PISA cycle. Missing values are not interpolated.
The vertical axis is limited to 440–520 to keep multi-cycle movement legible; PISA score scales extend beyond this displayed range.
United States mean scores by PISA cycle
A dash means there is no comparable U.S. score for that domain and cycle.
Scroll horizontally to view all columns.
| PISA cycle | Science | Mathematics | Reading |
|---|---|---|---|
| 2000 | — | — | 504 |
| 2003 | — | 483 | 495 |
| 2006 | 489 | 474 | — |
| 2009 | 502 | 487 | 500 |
| 2012 | 497 | 481 | 498 |
| 2015 | 496 | 470 | 497 |
| 2018 | 502 | 478 | 505 |
| 2022 | 499 | 465 | 504 |
| 2025 | 502 | 463 | 490 |
The United States has no published PISA 2006 reading mean because a U.S. test-booklet printing error affected reading instructions enough for the OECD to exclude the result. Mathematics and science results for 2006 were retained.
The three domains now show different long-run shapes
- Reading: the rounded mean stayed in a narrow 495–505 band from 2000 through 2022, apart from the missing 2006 result, then fell to 490 in 2025. The 2025 reading score is the lowest published U.S. reading mean in the PISA series.
- Mathematics: the series moved from 483 in 2003 to 463 in 2025. The 2025 result is about the same statistically as 2022, but both recent cycles are below the 2018 mean of 478.
- Science: the rounded mean moved from 489 in 2006 to 502 in 2025. Despite the higher endpoint, the OECD describes U.S. science performance as stable across PISA assessments, with the observed fluctuations not statistically different from the latest result.
Absolute Scores and the OECD Benchmark Do Not Move in Lockstep
A country’s own score trend and its distance from the OECD average are separate measures. The OECD average also changed between 2022 and 2025.
| Domain | U.S. 2022 | OECD 2022 | Gap 2022 | U.S. 2025 | OECD 2025 | Gap 2025 |
|---|---|---|---|---|---|---|
| Science | 499 | 485 | +14 | 502 | 482 | +20 |
| Mathematics | 465 | 472 | −7 | 463 | 463 | 0 |
| Reading | 504 | 476 | +28 | 490 | 461 | +29 |
Gaps are calculated from the published rounded means and are descriptive. Statistical comparisons should follow the OECD’s uncertainty tests rather than arithmetic gaps alone.
Reading illustrates the distinction clearly: the U.S. mean fell from 504 to 490, while the rounded U.S.–OECD gap stayed almost unchanged at about 28–29 points because the OECD reading average also fell. In mathematics, the U.S. mean was lower in 2025 than in 2018, yet it matched the 2025 OECD average because the OECD benchmark had also moved down.
PISA 2025 Proficiency Shares
Mean scores summarize a distribution. Proficiency shares show how many students reached the baseline level and how many reached Levels 5 or 6.
| Domain | U.S. at Level 2 or higher | OECD at Level 2 or higher | U.S. at Levels 5–6 | OECD at Levels 5–6 |
|---|---|---|---|---|
| Science | 78% | 74% | 13% | 7% |
| Mathematics | 65% | 65% | 8% | 8% |
| Reading | 74% | 69% | 13% | 6% |
Percentages are rounded as reported by the OECD. A system mean should not be read as the proficiency level of an individual student.
PISA 2025 also introduced computational problem solving within Learning in the Digital World. The United States recorded a mean of 513, compared with an OECD average of 500. That result is not part of the 2000–2025 core-domain trend because it does not have an equivalent historical series.
Participation, Sampling, and Trend Limits
The OECD marks the United States with an asterisk in PISA 2025 because one or more sampling standards were not met. In the United States, 4,625 students in 144 schools completed the assessment. NCES reports a 45.0% weighted school response rate before replacement, 53.9% after replacement, and a 76% student response rate. The OECD states that bias cannot be ruled out and that its likely direction cannot be determined.
More than 40% of responding U.S. students were not administered the student questionnaire in 2025. The OECD country note therefore limits questionnaire-based reporting. This page focuses on performance, proficiency, participation, and comparable trend measures rather than using incomplete questionnaire variables to explain score movements.
- Reading trend start
- 2000, when reading first served as a major PISA domain.
- Mathematics trend start
- 2003, after the mathematics scale was established for long-term comparison.
- Science trend start
- 2006, when science became the major domain and the comparable science scale was established.
- 2022 cycle timing
- The cycle originally planned for 2021 was administered in 2022 after a one-year delay.
- 2025 linking
- Trend tests include uncertainty from linking PISA 2025 scores to earlier cycles. This is one reason a visible point difference is not automatically a statistically different result.
Data Sources and Measurement
The score series uses official OECD PISA publications and U.S. technical documentation. Values shown in tables are rounded mean scores or percentages as published.
Official 2025 U.S. scores, proficiency shares, trend statements, participation counts, and the OECD’s data-quality caution. OECD — United States country note
Official international results, OECD averages, country/economy comparisons, and 2022-to-2025 changes. OECD — Student performance in PISA 2025
Official cycle-by-cycle U.S. means for reading, mathematics, and science through 2022, including the missing 2006 U.S. reading result. OECD — Overview of performance trends
Official PISA 2025 data files, documentation, and code resources. OECD — PISA 2025 Database
U.S. sampling, response rates, scaling, proficiency thresholds, linking uncertainty, and reporting limits. NCES — PISA 2025 Methodology and Technical Notes
OECD explanation of why the 2006 U.S. reading result was excluded after a printing error affected reading instructions. OECD — PISA Frequently Asked Questions