U.S. PISA 2025 Computational Problem-Solving Scores
U.S. 15-year-old students averaged 513 points in PISA 2025 computational problem solving, compared with an OECD average of 500 points. PISA 2025 was the first cycle to report this new computational problem-solving scale within Learning in the Digital World.
U.S. PISA 2025 Computational Problem-Solving Results
United States vs. OECD Average
Mean computational problem-solving performance among 15-year-old students in PISA 2025.
The United States is marked with an asterisk in OECD reporting because one or more PISA sampling standards were not met. The estimate should therefore be interpreted with added caution.
OECD PISA 2025 Results, Volume I, computational problem-solving performance tables.
Full U.S. and OECD Comparison
| Statistical Comparison | |||
|---|---|---|---|
| United States* | 513 | +13 | Higher than OECD average |
| OECD average | 500 | 0 | Reference average |
* OECD attaches a caution flag to the United States because one or more PISA sampling standards were not met.
U.S. Performance Was Above the OECD Average
- The 13-point difference between the U.S. mean and the OECD average is more than a numerical gap. OECD reporting identifies U.S. computational problem-solving performance as statistically higher than the OECD average.
- An exact ordinal position would give more precision than the data support. OECD identifies the United Kingdom, Switzerland, Czechia, Lithuania, Ireland, Austria, Belgium, Latvia, Poland, Finland, the Netherlands, Denmark, Luxembourg and Sweden among systems whose mean scores were not statistically different from the U.S. result.
- Computational problem solving is a new PISA 2025 scale. A 2022-to-2025 change cannot be calculated, and earlier PISA problem-solving assessments should not be combined with this scale as though they formed one historical series.
- The OECD sampling caution attached to the U.S. estimate remains relevant when comparing the country with other education systems, even though the reported U.S.–OECD difference passed OECD statistical testing.
Relative CPS Performance
This measure compares computational problem-solving performance with the level expected for students with similar science performance. The U.S. and OECD percentages are nearly identical numerically; the 0.1 percentage-point difference should not be treated as a statistical comparison.
What PISA Computational Problem Solving Measures
PISA 2025 introduced computational problem solving as the performance scale for the cognitive component of Learning in the Digital World. The assessment examines how students make progress toward learning and problem-solving goals while using computational tools and practices.
Tasks involve areas such as computer modelling, block-based programming, controlled experimentation, data analysis, and the construction or debugging of computational artefacts. Students work through extended digital tasks and can use resources such as tutorials, examples and feedback during the assessment.
This makes the scale different from a conventional test of programming knowledge alone. Performance reflects the use of computational tools to investigate problems, revise solutions and build functioning digital products within an interactive environment.
PISA 2025 Computational Problem-Solving Proficiency Scale
The computational problem-solving scale ranges from below Level 1 through Level 6. OECD does not define a formal baseline proficiency threshold for this domain. Level 3 is used as a useful reference point because it marks the stage at which students begin to demonstrate a broader set of functional computational problem-solving skills.
| Score Range | ||
|---|---|---|
| Below Level 1 | — | Below 330 |
| Level 1 | 330 | 330 to less than 427 |
| Level 2 | 427 | 427 to less than 468 |
| Level 3 | 468 | 468 to less than 512 |
| Level 4 | 512 | 512 to less than 566 |
| Level 5 | 566 | 566 to less than 622 |
| Level 6 | 622 | 622 to 1,000 |
The national mean of 513 points lies numerically just above the Level 4 cut point, but proficiency levels classify student performance distributions rather than assigning a proficiency category to a national mean.
How U.S. Performance Fits the International Distribution
OECD reported a wide range of computational problem-solving performance across participating education systems. Macao (China) recorded 572 points, Singapore 563, B-S-J-Z (China) 560 and Japan 557. Several OECD systems also recorded means above 530 points, including Estonia, Korea, Australia and New Zealand.
The United States was part of a broad cluster of education systems for which pairwise differences were not statistically clear. This is why the 513-point result should be read with its comparison intervals rather than converted into a simple international rank.
Across OECD countries, the average was 500 points. OECD also reported that 26% of students across OECD countries reached Levels 5 or 6, 21% reached Level 4, 17% reached Level 3, 13% reached Level 2, 17% reached Level 1 and 5% performed below Level 1. These percentages describe the OECD-wide distribution and should not be treated as the U.S. proficiency distribution.
Why There Is No PISA Computational Problem-Solving Trend Yet
PISA 2025 is the first assessment cycle using this computational problem-solving scale. The result therefore establishes a new measurement point rather than extending an existing U.S. trend line.
Earlier PISA cycles included other forms of problem-solving assessment, while other international studies use constructs such as computational thinking or computer and information literacy. Those measures have different assessment designs, scales and populations. Combining them with PISA 2025 computational problem solving would create a false time series.
Data Sources, Measurement and Comparability
OECD PISA 2025 Results
The U.S. mean score, OECD average, international comparison groups and computational problem-solving scale information are reported in PISA 2025 Results (Volume I): Future-Ready Students, including Table I.2.4 and the underlying PISA 2025 database tables.
United States Country Note
The OECD United States country note reports that U.S. computational problem-solving performance was higher than the OECD average and documents the U.S. participating student and school sample.
NCES PISA 2025 Technical Notes
NCES documents the U.S. implementation, scaling methodology and computational problem-solving proficiency cut scores. The scale runs from 0 to 1,000 points and is estimated using plausible values derived from item-response modelling.
Relative Computational Problem-Solving Performance
OECD relative-performance analysis compares computational problem-solving results with the performance expected from students with similar science scores. In the United States, 56.4% of students performed above that expected level, compared with an OECD average of 56.3%.
OECD PISA 2025 — Relative Performance in Computational Problem Solving
U.S. sampling caution: OECD marks the United States with an asterisk in PISA 2025 reporting because one or more sampling standards were not met. OECD and NCES results remain published, but estimates involving the United States require added caution.
Cycle comparability: PISA 2025 computational problem solving is a new scale within Learning in the Digital World. No directly comparable U.S. score exists for PISA 2022 or earlier cycles.
Population: PISA measures 15-year-old students enrolled in school rather than a particular U.S. grade. In the United States, 4,625 students in 144 schools completed PISA 2025, representing about 3.58 million 15-year-old students according to OECD reporting.
OECD — PISA 2025 Results, Volume I