Table of Contents
PISA 2025 Reading Decline: 28 Points, and What Parents Can Do
The PISA 2025 reading decline hit 28 points in a decade. Which explanations the evidence supports, which are thin, and three habits that target the loss.
The PISA 2025 reading decline is 28 points across OECD countries between 2015 and 2025. At the OECD’s own conversion rate of roughly 20 points per school year, that is about a year and a half of reading development, gone. Reading is now the steepest decline of the three core subjects and the OECD calls the trend “particularly concerning because reading skills underpin learning across the curriculum.”
But the shape of the loss is more informative than the size. Kids did not get uniformly worse at reading. They got worse at long reading, and at the specific act of evaluating and connecting what they read. And the share of students who answer fast and wrong nearly doubled.
Key Takeaways
- OECD average reading fell from 489 (2015) to 461 (2025), a 28-point drop. Non-OECD participants fell 16 points.
- The losses concentrated on tasks with long texts and on “evaluating and reflecting,” not on locating a single fact.
- “Hasty readers,” who answer quickly and incorrectly, rose from almost 7% in 2018 to 11% in 2025. Accurate, fluent readers fell about seven percentage points.
- Performance declined more in the later sections of the test than the early ones, which points at attention and persistence rather than decoding skill.
- School closure length does not predict reading loss: Sweden had none and fell 21 points; Costa Rica closed about 20 weeks and did not change significantly.
What the PISA 2025 reading decline actually looks like inside the test
PISA reading measures whether students can “understand, use, evaluate, reflect on and engage with texts.” In the United States, NCES runs the assessment. The 2025 results, published September 8, 2026, put the OECD average at 461, the lowest in the program’s 25-year history. Reading peaked around 2012 and has fallen since.
Then the report does something unusual for an international assessment: it breaks the decline down by item type.
Long texts lost the most. “Within the PISA reading test, the most rapidly declining success rates were observed on tasks that used longer texts as stimuli.”
Higher-order processes lost more than lower-order ones. Between 2018 and 2022, declines were sharpest on “higher-level processes, such as evaluating and reflecting, which require careful reading, remembering and integrating multiple pieces of information, sometimes across several sources.” Tasks requiring students to locate a single piece of information were least affected.
The fluency items at the start show an attention story. Accurate, fluent readers fell about seven percentage points between 2018 and 2025. Hasty readers, defined as students who “gave fast and incorrect responses,” rose from almost 7% to 11%. Meanwhile the share of items left unanswered fell, which the report reads as students becoming “more inclined to answer quickly rather than to leave difficult items unanswered.”
Fatigue shows up across the test. “Performance declined throughout the reading assessment, but more strongly in the later sections of the test than in the early sections.”
Put those four together and you get a specific picture. This is not a decoding crisis. It is a sustained-attention-on-demanding-text crisis. If you want the foundational-skills side of the story, our guide to the reading comprehension gap in kids who can decode covers where that breaks down earlier.
Hypotheses versus evidence strength
The OECD explicitly frames its four levels of explanation as “complementary” rather than rival, drawing on TALIS 2024 teacher data alongside PISA, and it does not rank them. The strength ratings in the last column are my reading of how well each is supported in this report, and I have tried to be conservative.
| Hypothesis | What the data actually show | Evidence strength |
|---|---|---|
| Attention and persistence declined | Hasty responses up, non-reached items down, bigger losses late in the test, self-reported effort down vs. 2018 and 2022 | Strong within PISA |
| Engagement with school declined alongside | Curiosity change correlates with reading change across systems (r = 0.54); “school is a waste of time” change correlates too (r = 0.53) | Moderate (correlational) |
| Classroom disruption ate learning time | TALIS 2024: time spent keeping order rose 13% to 16% between 2018 and 2024; threats between students correlate with reading declines (r = −0.63) | Moderate |
| Reading habits shifted to short digital text | More students report 1+ hour/day on social or games; fewer create content or seek information; 2023 meta-analysis finds digital leisure reading barely relates to comprehension (r = .055) | Moderate |
| Teacher support weakened | Fewer students report science teachers “continue teaching until students understand” (2015–2025), correlated with science change (r = 0.32) | Weak to moderate |
| Pandemic school closures caused it | Closure length does not predict reading loss; Sweden (no closures) −21, Costa Rica (~20 weeks) no significant change | Weak as a sole cause |
| Phones specifically caused it | PISA measures distraction (28% report classmates distracted in most science lessons) but offers no causal estimate of phones on reading scores | Thin |
| Changing student demographics caused it | ”PISA data show that changes in student body composition do not explain the overall decline” | Ruled out |
The honest summary is the OECD’s own: “The evidence does not point to one simple explanation for the performance decline. A more plausible interpretation is that performance declines reflect changes occurring at many levels, with small forces often aligning to produce strong effects.”
The phone question, handled fairly
Every parent wants the phone answer, so let me be precise about what can and cannot be said.
What PISA establishes: 28% of OECD students report classmates distracted by digital devices in most or every science lesson. In B-S-J-Z (China), Japan, and Korea, fewer than 10% do. Schools with clear mobile phone policies, “whether through bans or well-defined guidelines,” have fewer distracted students. More students in 2025 than 2022 report spending over an hour a day on social networks or games.
What PISA does not establish: that phones caused the reading decline, or by how much. It is an association study with no experimental arm on device access.
What outside research adds: the Altamura, Vargas, and Salmerón meta-analysis in Review of Educational Research (2023), covering 40 effect sizes and about 470,000 participants, found leisure digital reading correlates only weakly with comprehension (r = .055), against medium-sized effects for print reading habits. PISA 2018 data also showed students who read books on paper outperformed those reading mainly on devices, after accounting for socio-economic profile.
So: the direction of evidence is consistent with the phone hypothesis and cannot prove it. If you are looking for a decision rule, the print-versus-digital comparison has better support than the phone-ban claim. Our deep dive on print versus digital reading and comprehension works through that literature.
One finding I did not expect
Family conversation declined. The OECD reports that in a number of systems, “fewer conversations with parents or other family members were taking place at home to support students in their school life in 2025, compared to just three years earlier.” Specifically, fewer students reported that their parents asked them what they had done at school that day.
The report is careful to note this trend appeared equally in countries with large reading declines and in countries where reading held steady, so it does not explain the drop. I include it because it is the cheapest variable on the entire list, and it happens to be the one parents fully control.
What to actually do at home
The losses were concentrated on long texts, on integration across sources, and on sustained attention. Three habits target exactly those.
One long read a week, in one sitting
Pick something longer than your child would choose: a magazine feature, a chapter, a long article. One sitting, no device in reach. This is the direct counter to the finding that long-text tasks declined fastest. Length is the active ingredient, not genre.
One question that forces integration
After the read, ask one question that cannot be answered from a single paragraph. “How does what the author said at the start change what they claim at the end?” That is the “evaluating and reflecting” process PISA says fell hardest. One question is enough. Do not turn it into a quiz.
Make wrong-and-fast cost something
Hasty readers nearly doubled. When your child answers instantly and incorrectly, do not correct and move on. Ask them to point to the words in the text that support their answer. When they cannot, the lesson lands without a lecture.
Read where your kid can see you
PISA analyses have found students whose parents were seen reading at home, or who reported enjoying reading, tended to report more enjoyment themselves. The 2025 report also notes that reading lost ground as a favorite hobby among parents, not just children. This is not a guilt trip; it is a lever.
What not to do
Do not replace print with audiobooks and assume it is equivalent for this specific skill. Audiobooks have real benefits, and our piece on audiobooks versus reading comprehension covers them, but the skill PISA measured as declining is sustained visual attention to a long text with backward reference. And do not confiscate the phone as the whole plan. Removing distraction creates room for the long read; it does not produce one.
What to Watch For Over the Next 3 Months
- Week 4: Your child completes one long read in a single sitting without being reminded, even if it takes two tries to get there.
- Month 2 red flags: Answers given before the question finishes, or refusal to re-open a text to check something. Both are the hasty-reader pattern.
- Month 3 self-check: Can your child answer a question that requires connecting the beginning and end of a piece? That is the higher-order process that fell furthest.
Frequently Asked Questions
How big is the PISA 2025 reading decline in real terms?
The OECD average fell from 489 in 2015 to 461 in 2025, a 28-point drop. Using the OECD’s estimate that 15-year-olds gain about 20 score points per year of schooling, that is roughly a year and a half of reading development. The 2022-to-2025 portion alone was 14 points.
Did every country’s reading scores fall?
No. Türkiye rose 16 points in reading since 2022, the Slovak Republic 1, Costa Rica 1, and the United Kingdom held flat. Among non-OECD participants, Jordan (+20), the Philippines (+20), Cambodia (+18), and Georgia (+10) rose. Non-OECD participants fell 16 points on average versus 28 for OECD countries.
Is it screens or is it schools?
Both explanations have partial support and neither is established as the cause. The OECD’s position is that “small forces often aligning” produced the effect, spanning student attention, classroom disruption, system capacity, and societal reading habits. Anyone offering you a single cause is going beyond the data.
What about the pandemic?
It is part of the context but not a sufficient explanation. Reading peaked around 2012, well before COVID. And closure length does not predict the damage: Sweden closed no schools and lost about 21 points, while Costa Rica closed roughly 20 weeks and showed no significant change.
Do audiobooks count for building this skill?
They build vocabulary and listening comprehension, which matter. But the specific capability PISA found declining is sustained attention to long written text with integration across parts, which requires eyes on a page and the ability to look back. Use both; do not substitute.
About the author
Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.
Sources
- OECD. (2026). “PISA 2025 Results (Volume I): Future-Ready Students,” Box I.2.3 on drivers of the reading decline. OECD Publishing. https://www.oecd.org/en/publications/pisa-2025-results-volume-i_73451bc5-en.html
- OECD. (2026). “PISA 2025: Students’ reading and mathematics performance declined sharply across the OECD.” OECD press release, September 8, 2026. https://www.oecd.org/en/about/news/press-releases/2026/09/pisa-2025-students-reading-and-mathematics-performance-declined-sharply-across-the-oecd.html
- Altamura, L., Vargas, C., & Salmerón, L. (2023). “Do New Forms of Reading Pay Off? A Meta-Analysis on the Relationship Between Leisure Digital Reading Habits and Text Comprehension.” Review of Educational Research. https://doi.org/10.3102/00346543231216463
- OECD. (2025). “TALIS 2024 Results.” OECD Publishing. https://www.oecd.org/en/about/programmes/talis.html
- OECD. (2021). “21st-Century Readers: Developing Literacy Skills in a Digital World.” OECD Publishing. https://doi.org/10.1787/a83d84cb-en
- National Center for Education Statistics. (2026). “Program for International Student Assessment (PISA).” NCES. https://nces.ed.gov/surveys/pisa/
- Hong, J. (2026, September 11). “Students are using AI in school and test scores are down.” Fortune. https://fortune.com/2026/09/11/students-ai-school-test-scores-moderate-use-vs-daily-users/
