Socioeconomic Gaps in U.S. PISA Performance 2025
PISA 2025 reported a 104-point difference in U.S. science performance between students in the top and bottom national quarters of socioeconomic status. The corresponding OECD-average gap was 85 points. U.S. questionnaire-based estimates require particular caution because of coverage limitations documented by the OECD.
U.S. Socioeconomic Gap Indicators
Science Performance Gap by Socioeconomic Status
Score difference between the top and bottom national quarters of the PISA index of economic, social and cultural status
A positive gap means that students in the top national ESCS quarter had a higher mean science score than students in the bottom quarter. The U.S. questionnaire estimates carry an OECD coverage warning.
OECD PISA 2025 Results, Volume I, Table I.2 and Table I.B1.2b.2.
PISA 2025 Socioeconomic Indicators for the United States and OECD Average
| Reporting Note | |||||||
|---|---|---|---|---|---|---|---|
| United States* | 104 points | 11.0% | 10.8% | +6 points | +2 points | −4 points | Particular caution is required for questionnaire-based estimates because of U.S. coverage limitations. |
| OECD average | 85 points | 11.6% | 11.9% | +5 points | −8 points | −13 points | OECD average as published for the relevant PISA 2025 indicators. |
The U.S. Quartile Gap Was Wider Than the OECD Average
- The 104-point U.S. science difference between the highest and lowest national ESCS quarters was 19 points wider numerically than the OECD-average difference of 85 points. This measure compares two ends of the socioeconomic distribution rather than estimating the effect of a one-unit change in ESCS.
- The pattern looks different when socioeconomic association is measured as the share of score variance associated with ESCS. The U.S. estimate was 11.0%, close to the OECD-average value of 11.6%. A wide top-to-bottom quartile gap can therefore coexist with a gradient-strength estimate near the OECD average because the two indicators measure different features of the distribution.
- Among U.S. students classified as socioeconomically disadvantaged, 10.8% were academically resilient in science. The OECD average was 11.9%. Academic resilience identifies disadvantaged students who nevertheless scored in the top quarter of science performance within their own country or economy.
- The reported U.S. changes from 2022 to 2025 were +6 points among disadvantaged students and +2 points among advantaged students, producing a four-point numerical reduction in the gap. These rounded changes should be interpreted with the OECD’s U.S. data-quality warning and should not be treated as evidence of a reliable trend from their size alone.
What the PISA Socioeconomic Measure Represents
PISA measures socioeconomic status with the index of economic, social and cultural status (ESCS). The index combines information related to parental education, parental occupational status and household possessions associated with material and cultural resources.
For the national-quarter comparison, a socioeconomically disadvantaged student is in the bottom quarter of the ESCS distribution within the United States. An advantaged student is in the top quarter. The 104-point figure therefore describes the difference in mean science performance between these two U.S. groups; it is not the difference between fixed income categories and it should not be interpreted as a causal effect of family resources.
The variance measure answers a different question. The U.S. value of 11.0% represents the proportion of variation in science performance associated with differences in ESCS in the OECD model. It does not mean that socioeconomic status determines 11% of an individual student’s score.
Science Was the Main Socioeconomic Comparison in PISA 2025
PISA 2025 placed science at the center of its main performance reporting. OECD Volume I therefore presents the principal country-level socioeconomic snapshot using science performance, including the strength of the ESCS gradient, the advantaged–disadvantaged score difference and the share of academically resilient disadvantaged students.
Across OECD countries, the corresponding national-quarter gaps were 85 points in science, 77 points in reading and 83 points in mathematics. These OECD-wide values show that socioeconomic differences were present across all three traditional PISA domains. The U.S. figures on this page retain science as the principal comparison because the published U.S. socioeconomic snapshot is questionnaire-based and carries a specific coverage warning.
Data Sources, Measurement and Comparability
Official Results
The socioeconomic indicators are from OECD PISA 2025 Results, Volume I, particularly Table I.2 and the underlying PISA 2025 database tables.
Assessment Cycle
PISA 2025 assessed 15-year-old students. Results were released by the OECD on September 8, 2026. The assessment cycle remains 2025 even though publication occurred in 2026.
Socioeconomic Groups
Advantaged and disadvantaged students are defined using the top and bottom national quarters of the PISA ESCS distribution. These are relative socioeconomic groups within each participating education system.
Academic Resilience
An academically resilient student is in the bottom national quarter of ESCS but reaches the top national quarter of science performance within the same country or economy.
U.S. Sampling and Questionnaire Coverage
OECD reports a 45% U.S. school response rate before replacement and 54% after replacement, a 76% student response rate and a 6.9% overall exclusion rate. More than 40% of responding U.S. students did not receive the student questionnaire. Detailed reporting conditions are documented in the PISA 2025 Reader’s Guide.
U.S. Country Note
The OECD PISA 2025 United States country note states that bias in U.S. performance results cannot be ruled out and that its likely direction cannot be determined. Questionnaire-based indicators are omitted from that country note because of the coverage issue.
Database
Underlying student, school and questionnaire data are available through the OECD PISA 2025 Database.
Comparison Rule
Numerical differences do not by themselves establish that two estimates are statistically different. Sampling and measurement uncertainty must be considered when interpreting country differences and changes across cycles.