International assessments like PISA often seem like objective yardsticks: how well is our education system doing compared to others? But what if not all students put in the same amount of effort? And what if that effort varies between countries?

A new large-scale Swedish study spotlights that question. The researchers (Borger et al., 2025) combined PISA 2018 results from over 5,000 15-year-olds with their scores on Sweden’s national, high-stakes exams. The national tests count toward final grades and school progression. PISA, on the other hand, has no consequences for the individual student—a textbook example of a low-stakes test.

The central question: how much does a PISA score reflect a student’s ability, and how much does it reflect their willingness to try?

The results are apparent. Even when controlling for students’ proven ability on the national test, self-reported effort during PISA still predicted performance. More importantly, the link between the national and PISA results was significantly stronger for students who reported higher effort. In other words, PISA scores are a more accurate measure of ability for students who actually tried.

And the difference is far from trivial. On average, students who reported low effort scored about one-third of a school year lower than equally capable peers who put in more effort. Specifically, for low-effort students, each extra grade point on the national test translated into just +12 points on the PISA scale. For high-effort students, that became +20 points. Since 20 PISA points roughly equal one year of learning, effort alone can shift the picture from stagnation to progress. That’s not a footnote — it’s a structural distortion.

This is an uncomfortable message for policymakers who treat PISA results as direct reflections of educational quality. The study shows that PISA doesn’t just measure subject knowledge—it also captures test motivation. More specifically, it asks whether students are willing to exert effort on a test that offers them no personal reward. That’s not a flaw in the test itself, but it affects how we interpret the scores.

That doesn’t mean international comparisons are meaningless. On the contrary, they can offer valuable insights, especially when considered over time and in conjunction with other assessments. If PIRLS, PISA, TIMSS, and national data all signal a problem, then yes, there likely is a real problem. But it’s important to realise that cross-country score differences may also reflect differences in test motivation, not just between students, but also between countries. In some places, students may take these tests more seriously than others. And that can affect the outcome.

This study is especially compelling because it doesn’t stop at surface-level correlations. Robust statistical modelling disentangles the role of effort from ability. That effect is small but significant — and, crucially, practically relevant.

Perhaps the most important insight of all is that because PISA is a low-stakes test, it may tell us something else, too. It reveals how young people act in situations where their individual effort isn’t rewarded—how willing they are to take responsibility even when no one’s watching. That kind of motivation might be something we need to take much more seriously in education.

Abstract of the study:

It is commonly recognized that test performance is influenced by both cognitive ability and motivational factors. To explore this phenomenon, a random sample of 15-year-old Swedish students (n = 5504), whose PISA 2018 results were linked to national registry data, was analyzed. Students’ PISA performance was regressed on their self-reported test-taking effort in PISA, their national test scores in corresponding subject domains, and the interaction between these variables. Results reveal that test-taking effort had an independent influence on PISA scores after controlling for high-stakes test results (β = 0.15). More importantly, the relationship between the high-stakes national test and the low-stakes PISA was stronger at higher levels of effort (interaction coefficient; β = 0.05). Students who report low effort underperform in PISA by approximately one-third of a school year’s learning gain, emphasizing the role of effort in obtaining an accurate assessment of ability. Implications for research and practice are discussed.