How Türkiye Became One of the Fastest-Improving Countries in PISA: 2022–2025

Türkiye entered PISA in 2003 from a position well below the OECD average in the core subjects. Two decades later, its 2025 results place the country above the OECD mean in science and reading and almost level with it in mathematics. Türkiye scored 494 in science, 472 in reading, and 462 in mathematics, compared with OECD averages of 482, 461, and 463. The pattern matters because Türkiye did not merely move upward in a ranking: its own scores rose while OECD averages moved down in all three subjects between 2022 and 2025. [OECD]

The latest cycle produced an unusually strong three-subject gain. OECD trend estimates show Türkiye rising by 18 points in science, 16 in reading, and 8 in mathematics from 2022 to 2025. In OECD Education GPS, the reading increase ranks fourth among 50 systems with comparable data, the mathematics increase seventh among 46, and the science increase ninth among 27. That is a firmer basis for describing Türkiye as one of PISA’s fastest-improving systems than rank movement alone. [OECD]

PISA 2025 was computer-based in Türkiye and ran from April 14 to May 16, 2025. Science was the main assessment domain. The national sample included 7,702 students in 203 schools across 56 provinces. [MEB]

There is a small methodological detail behind the mathematics change. The rounded published means are 453 in 2022 and 462 in 2025, which appear nine points apart. OECD’s official change estimate is +8 points because trend calculations use the underlying unrounded estimates. The two figures therefore describe the same result at different levels of precision. [OECD]

The 2022–2025 Jump Was Larger than Rankings Alone Suggest

Türkiye’s rank gains were large, but rankings are not the cleanest measure of change because the PISA participant pool also changed. PISA 2022 included 81 countries and economies; PISA 2025 included 91. A rank can move because a country’s score changes, because other systems move, or because the participant list changes. The underlying score trend is therefore the better starting point.

Among OECD members, Türkiye moved from 29th to 14th in science, from 30th to 13th in reading, and from 32nd to 23rd in mathematics. Those rank changes are noteworthy, yet the stronger evidence is that Türkiye’s mean performance rose in all three subjects. Exact PISA positions should also be read with care because scores are sample estimates and nearby systems can have overlapping uncertainty ranges. OECD identifies Türkiye as one of the few participating systems to improve across science, reading, and mathematics between 2022 and 2025. [OECD]

Türkiye Rose While OECD Averages Fell

The 2022–2025 period becomes more informative when Türkiye is compared with the OECD mean. In 2022, OECD averages were 485 in science, 476 in reading, and 472 in mathematics. By 2025 they were 482, 461, and 463. Türkiye moved in the opposite direction in every subject. [OECD]

1. Science: Türkiye moved from 9 points below the OECD mean in 2022 to 12 points above it in 2025.
2. Reading: the gap moved from 20 points below the OECD mean to 11 points above it.
3. Mathematics: the gap narrowed from 19 points to just 1 point.

This distinction answers an important question about the rise. Part of Türkiye’s rank improvement reflects weaker OECD-wide results, especially in reading and mathematics. Yet the improvement cannot be explained by other countries falling alone. Türkiye’s own scores increased at the same time. In reading, that combination produced a 31-point swing relative to the OECD average between 2022 and 2025.

The Long-Run Trend Started Well Before 2025

Türkiye’s 2025 surge is easier to understand as the latest stage of a much longer movement. Mathematics rose from 423 in 2003 to 462 in 2025, while reading rose from 441 to 472. For science, OECD’s comparable trend begins in 2006, when science first served as PISA’s main assessment domain; on that basis Türkiye rose from 424 in 2006 to 494 in 2025. [OECD]

MEB also reports a 2003 science mean of 434. [MEB] That number is useful in national historical reporting, but it should not be silently joined to the OECD’s later science trend line. For a like-for-like science series, 2006 is the cleaner baseline. This is why the chart below begins all three lines in 2006 even though Türkiye first participated in PISA in 2003.

The series does not show uninterrupted progress. Mathematics and reading climbed strongly through 2012, all three domains dropped sharply in 2015, and performance rebounded in 2018. Science then continued upward in 2022 and accelerated again in 2025. Reading fell between 2018 and 2022 before recovering to 472. Mathematics was nearly flat between 2018 and 2022, then reached 462 in 2025. The long-run direction is upward, but the path is uneven.

Why 2015 Should Not Be Treated as the Long-Run Direction

PISA 2015 stands out from the rest of Türkiye’s series. From 2012 to 2015, mathematics fell from 448 to 420, reading from 475 to 428, and science from 463 to 425. By 2018, the scores had returned to 454, 466, and 468. OECD’s trend analysis treats the 2015 results as an unusual break rather than the direction of the longer series. [OECD]

OECD discusses several possible explanations without treating any of them as proven. PISA moved from paper-based testing in 2012 to computer-based testing in 2015, so differences in familiarity with digital testing may have affected results. OECD also notes changes in sampling as the distribution of students across school types changed. The organization explicitly describes the sampling explanation as speculative without further examination. [OECD]

This matters for 2025 because it prevents two misleading readings of the history. The 2012–2015 fall should not be treated as a normal three-year deterioration, and the 2015–2018 rebound should not be presented as if all of it represented newly created learning. The cleaner story is the multi-cycle trend.

Access Expanded at the Same Time as Scores Improved

One of the less visible parts of Türkiye’s PISA story is that the measured student population became much broader. Between 2003 and 2018, the number of 15-year-olds eligible for PISA increased by more than 400,000, while PISA sample coverage of the 15-year-old population more than doubled from about 36% to 73%. [OECD]

That expansion changes how the score trend should be interpreted. When an education system keeps more adolescents in school, students who previously would have been outside the measured school population enter the assessment. That can place downward pressure on the national average. Türkiye instead managed to expand participation while improving performance. OECD notes that in reading, both the 25th and 75th percentiles improved by 27 points between 2003 and 2018, suggesting that progress was not confined to students already near the top of the distribution. [OECD]

The timing also matters. OECD records a progressive rise in enrollment among 15-year-olds through 2012, with the largest rise when education became compulsory for this age group. The long-run improvement therefore occurred during a period when the secondary-school population itself was changing, not while PISA measured a fixed and selective group.

The Rise Reached Both Baseline and Advanced Performance

A higher national mean can come from different parts of the score distribution. PISA 2025 shows that Türkiye’s gains were visible among students reaching the minimum proficiency threshold and among high performers. In science, 81% reached at least Level 2, compared with 74% across the OECD. Reading reached 76% versus 69%. Mathematics was almost identical to the OECD average at 64% versus 65%. [OECD]

The longer change at the lower end is even clearer. Compared with 2015, the share of students below Level 2 fell by 25 percentage points in science and by 16 points in both reading and mathematics. That pattern makes it difficult to explain the national rise as an outcome limited to a small group of highly selective schools or only to the highest-scoring students. [OECD]

The upper end improved as well. MEB reports that Türkiye’s Level 5–6 share in science rose from 4.0% in 2022 to 7.4% in 2025. In reading, the top-performer share rose from 1.9% to 4.2%; in mathematics, from 5.4% to 7.9%. [MEB]

What Changed in the Education System Before 2025

PISA cannot isolate one reform and assign it a share of a country’s score gain. Türkiye’s long record instead points to cumulative changes in access, curriculum, assessment practice, school capacity, and the composition of secondary education. The most defensible explanation is therefore system change accumulated over many years, not a single intervention immediately before the test.

A Longer Shift Toward Applied Skills

Türkiye began a broad curriculum change in 2005 that moved away from heavily prescribed content and greater reliance on knowledge transmission. Later revisions, including those in 2017, reduced content load and gave more weight to competencies such as problem solving, critical thinking, student self-assessment, projects, and varied assessment methods. Those changes align more closely with PISA’s emphasis on applying knowledge to unfamiliar problems than a curriculum centered mainly on recall. [OECD]

Broader Secondary Participation Changed the Student Base

Expansion of schooling is not just an access statistic; it changes who appears in an assessment of 15-year-olds. The rise from 36% to 73% PISA population coverage by 2018 means that later results represent a much broader slice of Türkiye’s adolescent population than early cycles did. Maintaining an upward score trend while the tested population becomes broader is harder than improving within a narrow, stable group.

Staffing Conditions Improved Between 2022 and 2025

In 2025, 8% of students in Türkiye attended schools whose principals said instruction was hindered to some extent by a lack of teaching staff, down from 16% in 2022. The OECD average in 2025 was 39%. This is one school-condition indicator moving in a favorable direction during the latest score rise. The resource picture is not uniform, however: 24% of students were in schools reporting that shortages of educational materials hindered instruction, up from 15% in 2022, while the physical-infrastructure measure was nearly unchanged at 14% versus 15%. [OECD]

What May Help Explain the 2022–2025 Acceleration

The latest jump is large enough to invite a search for one immediate cause, but PISA is not an experiment. It can show that conditions changed alongside scores; it cannot prove that one policy produced a particular number of PISA points. Several 2025 indicators are still useful because they describe the learning environment around the cohort.

• Perceived teaching-staff shortage fell: the share of students in affected schools dropped from 16% in 2022 to 8% in 2025.
• Peer digital distraction was below the OECD average: 18% of students said peers were often distracted by digital devices during most or all science lessons, compared with 28% across the OECD.
• Whole-school cell-phone bans became more common: 83% of students attended schools where phones were not allowed on the premises, up from 62% in 2022. The OECD average in 2025 was 49%.

These indicators are associations, not a causal decomposition of the score increase. The phone data illustrate the point well: students in Türkiye attending schools with a ban were less likely to report digital distraction, but that does not establish that phone rules produced the national PISA gain. School composition, teaching practice, student behavior, family background, and many other variables move at the same time. [OECD]

The 2025 Cohort Mostly Learned Under Earlier Curricula

The timing of Türkiye’s newest curriculum matters when interpreting PISA 2025. The Türkiye Yüzyılı Maarif Modeli began in the 2024–2025 school year on a phased basis in preschool, Grade 1, Grade 5, upper-secondary preparatory classes, and Grade 9. [MEB]

PISA testing took place in April and May 2025, only months after that phased introduction. Most 15-year-olds in the PISA cohort had completed nearly all of their schooling under earlier curricula. It is therefore too early to attribute most of the 2025 score jump to the new model. Its effect on PISA-age students is better judged in later cycles, after cohorts have experienced the curriculum for longer.

Socioeconomic Results Show Both Reach and Remaining Distance

PISA 2025 offers an unusually useful view of Türkiye’s performance among lower-income student groups. Thirty-two percent of students in Türkiye fell in the bottom international quartile of PISA’s economic, social, and cultural status scale. Their average science score was 466, which OECD describes as one of the highest scores among students with a similar socioeconomic background. [OECD]

Within Türkiye, the science gap between the most advantaged and least advantaged quarters was 81 points, close to the OECD average gap of 85. Some 13% of disadvantaged students reached the top quarter of science performance within Türkiye, compared with 12% across the OECD. At the same time, the advantaged–disadvantaged science gap widened between 2015 and 2025 because performance rose more among advantaged students. The national mean can therefore improve while distributional gaps still require separate attention.

A similar caution applies at the very top and bottom of the total score distribution. Between 2022 and 2025, OECD reports that the gap between the highest-scoring 10% and lowest-scoring 10% widened in mathematics and science, while reading showed no clear change. This does not erase the broad fall in low performance since 2015; it shows that average improvement and equal improvement are not the same measure. [OECD]

Attendance Data Tell a Separate Story

Strong test-score growth did not coincide with uniformly strong attendance indicators. In PISA 2025 student questionnaires, 53% of students in Türkiye reported skipping at least one full day of school in the two weeks before the assessment; the OECD average was 22%. Another 41% reported skipping at least one class, while 49% reported arriving late. The corresponding 2022 figures in Türkiye were 53%, 40%, and 45%, so the full-day absence measure was essentially unchanged across the two cycles. [OECD]

These are student self-reports, not administrative attendance records, so they should be interpreted as questionnaire indicators. Even with that limitation, they show that Türkiye’s PISA rise is not evidence that every school-process measure moved in the same direction. Learning outcomes and attendance need to be tracked separately.

Core Scores Rose Faster than Digital Problem-Solving Performance

PISA 2025 added a new assessment of computational problem solving, measuring how students use computational tools and practices to solve unfamiliar tasks. Türkiye averaged 473 points, below the OECD average of 500. This creates a different picture from science and reading, where Türkiye had moved above the OECD mean. [OECD]

Digital use itself is not low. In the same PISA cycle, 52% of students in Türkiye reported using AI chatbots for learning at least once a week, compared with 46% across the OECD. Forty-four percent reported using AI for preliminary research on a new topic, against an OECD average of 31%. The contrast suggests that frequent use of digital tools is not identical to high performance in computational problem solving. [OECD]

What PISA 2025 Changes About Türkiye’s Long-Run Story

The 2025 results strengthen a trend that was already visible before the latest cycle. Türkiye expanded the share of 15-year-olds represented in PISA, lifted long-run mathematics and reading performance from its early-2000s levels, raised science strongly from the comparable 2006 baseline, and reduced the share of students below Level 2 over the past decade. The 2022–2025 cycle then added an unusually fast three-subject rise.

Three mechanisms explain why the latest movement looks so large. First, Türkiye made absolute gains: its own science, reading, and mathematics scores rose. Second, it made relative gains because OECD averages fell over the same period. Third, the improvement was visible beyond the mean: baseline proficiency rose, while top-performer shares also increased. None of those points requires treating rank position as the sole measure of progress.

The next test is persistence. If later PISA cycles keep the 2025 score level while narrowing performance gaps, improving attendance indicators, and lifting computational problem-solving results, the 2025 jump will look less like an exceptional cycle and more like a durable change in Türkiye’s education outcomes.