The results of the 2025 PISA, a global study conducted by the Organization for Economic Co-operation and Development (OECD) that measures 15-year-olds’ academic performance in math, science and reading, were released Sept. 8. They showed Argentina posting its worst-ever score in mathematics and Mexico its lowest since 2006, while Brazil, Colombia and Costa Rica remained stagnant and Chile fell back in both math and reading. What factors explain the declines, and how much do the results stem from the Covid-19 pandemic’s lasting effects on learning? Why did some countries manage to hold their scores steady while others deteriorated sharply? What education policies could improve results?
Donna Hrinak, co-vice chair of the Inter-American Dialogue’s Board of Directors and director at Covista: “The performance of students in Latin America on the global PISA exams has long been shocking. Throughout this century, results for the region have been among the lowest in the world, generally just slightly above outcomes in Sub-Saharan Africa. Again this year, Latin America is considerably below the OECD average. The results of the latest 2025 PISA demand more than shock: They cry out for a sober consideration of how we can ensure young people in Latin America can compete and excel—and contribute to the success and excellence of the region’s economies and societies. What have we been doing wrong and what needs to change? Fortunately, the region has already made great progress in student attendance. We know getting kids into school and keeping them there is a critical first step. Now we need to take the next steps, not one simple formula, but a holistic approach that includes better teacher preparation and increased parental involvement with their kids’ schools. At the same time, it’s important to keep in mind that the scores of 15-year-olds are only a snapshot in the video of their entire education. Improving early childhood education might be the fastest way to reverse regional declines. And as I write this in the United States, I note that reading and math scores of students here have also fallen and are near their lowest point in 25 years. Like so much, improving educational outcomes is a shared interest across our hemisphere.”
Sarah Stanton, program manager, and Micaela Finoli, research associate, both at the Education Program of the Inter-American Dialogue: “Every three years, the release of PISA results produces recriminations, a few celebrations and attempts to read the tea leaves about the future of education. This year, the results provoked a particularly strong reaction globally, because it was not only the ‘usual suspects’ who obtained poor scores, but almost all countries saw a marked decline, even those often upheld as exemplars. In Latin America, where 13 countries participated, there was a broad but not homogeneous decrease in performance: Uruguay and Chile remained stable, while in El Salvador only 13 percent of students met minimum proficiency in mathematics, and Mexico and Argentina posted some of their worst results ever. The decline cannot only be attributed to the pandemic: in several countries, the downward trend predates 2020, although Covid closures were an accelerator. What is a global phenomenon is the impact of technology on learning, for good or ill: While moderate technology use for learning is associated with better performance, excessive use and digital distraction worsen performance; surprisingly, PISA results also show that students whose reading scores dropped the most were from wealthier families, and the percentage of ‘hurried readers’ has almost doubled since 2018. The central concern is not the technology itself, but rather its use. While PISA results are not designed to determine the effectiveness of any particular education policy, statistical correlations indicate that countries with better performance tend to prioritize less content but teach it in depth, strengthen teacher training, protect reading time and recover attendance, punctuality and a classroom climate that allows teaching, sustaining these reforms beyond a cycle of government.”
Alejandra Mizala, rector of the University of Chile: “Covid-19 undoubtedly matters: The students assessed in 2025 were in primary school when schools closed, and Chile experienced particularly prolonged closures. Yet learning outcomes had begun to stagnate or decline in many countries before the pandemic, suggesting that longer-term factors are also at play. Chile is an interesting case. It made substantial progress across all three PISA domains during the 2000s, followed by a long period of stagnation. In 2022, Chile appeared relatively resilient to the pandemic shock. In 2025, however, reading fell by 12 points and mathematics by nine, while science remained stable. This uneven pattern suggests that Covid alone cannot explain the results. We need to examine what is happening inside schools and classrooms, as well as how young people are learning. Classroom disruption, weaker student engagement and insufficient efforts to strengthen teachers’ capacities matter. Changes in how young people read and use technology may also be playing a role, particularly when digital devices and artificial intelligence distract students or replace, rather than support, their own thinking. There is no single explanation, and therefore no single solution. Differences across countries are likely to reflect, among other factors, their institutional capacity, the effectiveness of learning recovery policies and the conditions for teaching and learning. The policy response should therefore be equally multidimensional: strengthen teaching preparation and school leadership, improve classroom environments, provide targeted pedagogical support and systematically learn from schools that achieve strong results despite challenging circumstances.”
Santiago Cueto, principal researcher at the Group for the Analysis of Development (GRADE): “The recently released PISA results from 13 Latin American countries are disheartening as they reveal that, on average, nations in the region lag between three and five grade levels behind OECD countries. Also, test results are linked positively with students’ socioeconomic status, thereby highlighting significant educational inequity. While PISA assessments are designed to measure skills and associated factors, they do not strictly pinpoint the causes of low and unequal performance. Among these potential causes, first, the loss of instructional time during the pandemic cannot be ruled out, given the evidence that more class time correlates with better performance. Second, there has been considerable speculation regarding the role of excessive AI use in poor performance. Indeed, recent studies indicate that while unguided AI use may yield superior academic outputs, it ultimately may lead to a negative impact on students’ skills. Regarding potential policy responses, the literature on what works shows that interventions based on structured pedagogy, teaching at the right level, and fostering a growth mindset have a positive impact on learning, yet these approaches are rarely implemented in Latin America and the Caribbean. Furthermore, the 15-year-old students participating in PISA have educational histories suggesting that the roots of low performance can be predicted as early as the preschool years. The above is to suggest that any policies adopted considering PISA results should be grounded in recent empirical evidence—ideally generated within the region or in other developing nations.”
Claudia Romero, former director of the Education Department at Torcuato Di Tella University and creator of the UTDT School Leadership Program: “The 2025 PISA results show that Latin America is falling further behind in education. Understanding this situation requires considering its multiple causes, although several significant factors can be identified across the region. On the one hand, the ‘long shadow’ of the Covid-19 pandemic continues to affect learning outcomes, given that school closures in the region were among the longest in the world. Their impact has been compounded by pre-existing socioeconomic vulnerability. At the same time, the rapid post-pandemic adoption of technology and artificial intelligence tools available to students raises concerns that inappropriate use may substitute for attention, effort and comprehension, affecting academic performance. Alongside these factors, there is also a methodological mismatch between the competency-based educational model assessed by PISA and the reality of Latin American classrooms. Teaching practices remain largely guided by fragmented, content-based curriculums, despite curricular reforms that appear to have ‘remained on paper’ because the necessary working conditions and teacher training for their implementation have not been put in place. To address these factors, it is urgent to sustain and increase investment in education to reduce the impact of students’ socioeconomic vulnerability, implement measures to address learning deficits resulting from Covid-19, improve teachers’ working conditions to enable meaningful transformation and promote critical digital literacy.”
Nicolás Mariscal, chairman of Grupo Marhnos in Mexico City: “The PISA report revealed the sharpest global academic decline among OECD countries. Latin America showed a mixed landscape: Chile, Uruguay and Costa Rica displayed relative strength, while Argentina fell to its lowest level on record. Mexico maintained apparent stability in science, reading and mathematics, avoiding total collapse, yet it remained near the bottom of OECD rankings. Andreas Schleicher, the director responsible for the PISA report, attributed this resilience in Mexico to students’ socioemotional perseverance, as well as support from teachers and peers. President Claudia Sheinbaum endorsed this view to defend public education while proposing investments in infrastructure and teacher training. However, official budget data from Mexico’s Secretariat of Public Education (SEP) show significantly more resources allocated to scholarships than to targeted infrastructure. The Mexican Institute for Competitiveness (IMCO) warns that relative stability relies on family effort rather than effective public policy, pointing out that half of Mexican youth struggle with basic reading comprehension and scientific thinking. IMCO clarified that the problem is not a lack of student motivation, as eight out of 10 show a desire to learn, highlighting a failure in educational policy rather than in youth attitude. While the pandemic worsened global learning loss, the downward trend began in 2012. Evidence shows recreational screen time severely impaired reading comprehension, costing the equivalent of one school year. Ultimately, OECD and IMCO experts agree that overcoming this crisis requires reallocating top teachers to vulnerable schools, prioritizing early childhood education, restoring independent standardized assessments, and transitioning from memorization toward real-world problem-solving as a core component of the educational process.”
The Latin America Advisor‘s daily Q&A features expert analysis from leading policymakers, analysts, scholars and industry leaders to offer diverse perspectives on one of the region’s most pressing issues. It is available to members of the Dialogue’s Corporate Program and others by subscription.
