The reading scores of U.S. 15-year-olds on an international exam dropped significantly over the past three years—an amount exactly mirroring the decline among industrialized nations, according to international test scores released today.
The 14-point decline in reading on the Program for International Student Assessment is especially notable for its magnitude, which translates to about six or seven months’ learning, according to estimates by the Organization for Economic Cooperation and Development, which administers PISA. There is suggestive evidence from the exam that students are struggling in particular with longer, more complex texts.
The findings generally align with separate gauges of younger American students’ reading scores, which have shown continued declines in reading, even as math performance appears to have stabilized.
American students’ performance in math and science, the other two tested subjects on PISA, was statistically no different from 2022. But learning in those subjects has declined when viewed over the longer lens of the past decade—both in the United States and globally.
And the United States’ math performance is comparatively weaker than its reading or science performance when looking across countries.
The findings are almost certain to bring new attention to reading policy in the United States—as well as fueling red-hot concerns among policymakers and educators alike about the impact of students’ screen time and social media use, and the effect of the rapid adoption of generative AI technology on reading and writing habits.
The consistency in declines across other industrialized nations also suggests many are facing parallel challenges.
“We cannot say for sure, but declining levels of reading for enjoyment and the shift towards digital reading—skimming social media feeds and rapidly processing information—may be compromising the ability of students to engage with complex texts and data,” said Andreas Schleicher, the director for education and skills for the OECD, in a report accompanying the scores.
PISA does not offer cause-and-effect conclusions allowing researchers to definitively link such trends to the test scores. But the data are worrisome, said Matthew Soldner, the acting commissioner of the National Center for Education Statistics, part of the U.S. Department of Education.
He noted that, in addition to the reading declines, the United States now has the largest gap in reading performance between its lowest and highest-achieving students of any of the dozens of countries that participated in PISA.
“It just strikes me that what this shows is that the U.S. has the ability to educate students exceptionally well,“ he said. “It really indicates our struggling students are not getting the full benefit of what we know how to do this country. And the unevenness in how we are supporting students, educators, systems of education is not doing us any favors.”
Understanding how to interpret PISA results
The PISA tests were given in 2025 to some 760,000 students across in 91 countries and economies in all. The 38 OECD countries that participate in PISA are generally industrialized and wealthier, make them a better comparison point to the United States than the global average of all test-takers.
(New participants in the 2025 results included Armenia, the capital area of Tajikistan, Ecuador, Kenya, Rwanda, and Zambia.)
Each administration of PISA places special emphasis on one area. In 2025, that topic was science. American 15-year-old students’ science performance held steady, with their scores roughly comparable to those of students in the United Kingdom, Canada, New Zealand, Australia, and Finland, among other countries.
But as in reading, the gap between the United States’ highest- and lowest-performing groups of students, at nearly 300 points, was larger than that in most of the countries who took PISA, and in 12 out of the 13 countries whose scores in science were comparable. Of those countries, only Luxembourg had a higher gap between its strongest and weakest students.
The widening of this gap over the past decade has also emerged on the nation’s report card, a U.S.-only gauge of student performance.
U.S. Secretary of Education Linda McMahon cited the gap on the PISA to bolster her administration’s support of school choice.
“Today’s international education scores confirm a decade of stagnation and a devastating gap: We successfully propel our highest achievers, but the lowest-performing students are falling further behind,” she said in a statement. “This moment is a stress test for our nation’s future. To pass it, we must enact a hard reset that stops protecting a failed status quo and instead builds a system that empowers state leaders and embraces innovative learning options through school choice.”
Here’s a look at some of the main takeaways from the 2025 administration of PISA, which is given to 15-year-olds.
- In science, the main focus of this year’s PISA, the U.S. score did not change. Seven education systems had higher scores than the United States average: Four Chinese jurisdictions, as well as Singapore, Macao, Taiwan, Japan, Estonia, and Korea. The average score among OECD countries—mainly other industrialized nations—was lower than the U.S. score, which did not change from 2022.
- In reading, U.S. scores fell by the same amount, 14 points, as the average across other OECD countries. Still, it had among the highest performance in that subject and outperformed other OECD nations. Only three education systems outperformed it: Singapore, the Chinese jurisdictions, and Taiwan.
- In math, U.S. scores did not change from 2022. But its overall comparative ranking, at 27th, is lower than in the other two subjects.
PISA results often result in handwringing among politicians, and attempts by advocates and researchers to look for clues about what might be helping fuel success in the top nations. The 2025 scores are likely no exception, and they include some intriguing, if difficult-to-interpret findings.
For example, among predominately English-speaking nations, the United Kingdom’s reading scores did not change, while scores in Australia, Canada, and Ireland all declined significantly.
Finland, a darling of education policy analysts for its strong performance in the early 2000s, has fared disastrously on recent PISA administrations: Its reading and math scores, already in decline, fell by 16 and 15 points respectively in 2025. Turkey was the only OECD country that improved significantly in all three subjects.
But drawing conclusions about what fueled those changes can be tricky, especially because of the major differences in how countries structure their education systems and fund them. Many, for example, have a centralized curriculum or set of standards, unlike the United States’ diffuse, locally controlled system, which is mainly shaped by states and local school districts.
PISA is also unique in what it measures. Its test frameworks emphasize problem-solving and application over assessing grade-level content expectations, as on most U.S. tests. In science, PISA focuses heavily on constructing appropriate scientific investigations and evaluating data; in math, on solving problems in real-world contexts; and in reading, on using texts to further goals.
Could reading declines be linked to ‘hasty reading’?
PISA also released detailed background data that are likely to be mined for insights into the reading results. Some of them give insights into how students are using AI tools.
About half of students said they use AI once a week or more to help themselves learn, and one in three said they used it to draft text for writing assignments. The data show a generally negative correlation between student use of AI to aid writing, and reading score on PISA.
And 15-year-olds globally seem to be spending more time browsing social networks; the share of students who said they spent more than an hour each weekday doing that grew by about 6 percent.
On the reading test itself, students on PISA appeared to find specific kinds reading demands particularly challenging. They struggled with tasks that used longer texts and to evaluate information across multiple texts. And the data show a decline in fluency, or how quickly and accurately students read, combined with a higher proportion of “hasty readers” who consumed the texts rapidly and answered quickly and incorrectly.
Those factors “suggest students’ attention, persistence, or confidence when facing difficult tasks declined over time,” the report accompanying the scores said.
Those suggestive findings dovetail with concerns about older readers in the United States as well. Prior EdWeek reporting has shown that many educators of students in grades 3-8 are worried about their students’ inability to read complex texts at length. And a consensus is beginning to emerge that older readers who struggle to read need specific kinds of supports and interventions.
Few students in the United States are tested at age 15, so the PISA findings could fill in a gap showing that reading problems appear to persist into upper the upper secondary level. Many of the students who took the exam would have been still in elementary or middle school during the COVID-19 pandemic—a period in which students are generally expected to be reading longer and more difficult texts.
“It likely shows they developed foundational literacy skills, but they did have interruptions when they were at the next level, and that might be why we see some of this hangover,” said Soldner of NCES.
Reading has been a major focus of U.S. policy over the past decade. Most states have passed laws or other policy that including an emphasis on explicitly teaching alphabetics, including phonics, but far fewer states have designed comprehensive plans for older readers.
More recently, parents and educators alike have pushed back on the ubiquity of screens in classrooms. Amid this “techlash,” major school systems, including New York City and Los Angeles, have frozen or placed severe limitations on the use of AI in classrooms.
