5 Evidence-Based Strategies To Improve U.S. K-12 Math Achievement in Wake of 2025 PISA Results

It is no secret that academic achievement in the United States has declined, a trend that began in 2013, well before the COVID-19 pandemic temporarily closed schools to in-person learning and forced school systems to transition to remote instruction.¹ And although American students are experiencing declining performance in both reading and math,² conversations and legislative action have mainly centered around reading. However, prioritizing math performance is equally critical. Stronger math skills predict higher future earnings and are essential for in-demand jobs in manufacturing, health care, and technology—especially as artificial intelligence expands.³ With 56 percent of employers struggling to find skilled candidates, policymakers must invest in boosting K-12 math performance to ensure U.S. students are prepared for an evolving labor market.⁴

Recent results from the 2025 Programme for International Student Assessment (PISA), which evaluates 15-year-olds globally in math, reading, and science, raise renewed concerns about U.S. student performance.⁵

Between 2022 and 2025, achievement across all three domains either declined or remained stagnant.⁶ In math specifically, the 2025 U.S. average score of 463 hit an all-time low and was not statistically different from students’ 2022 performance, with only 8 percent of U.S. students qualifying as top performers.⁷ Most strikingly, the United States recorded one of the widest gender gaps among developed nations with boys outperforming girls by 24 points, nearly double the average gap of 13 points reported by the Organisation for Economic Co-operation and Development (OECD).⁸ This stark disparity highlights an urgent need to ensure girls achieve parity in foundational mathematics skills.

This report details the causes of declining math achievement in the United States and outlines successful state efforts to improve students’ math scores. Finally, it provides five policy recommendations to accelerate math achievement across the country. Policymakers should:

• Strengthen accountability and align academic standards
• Expand the use of high-quality instructional materials and high-impact practices
• Enhance teacher preparation and in-service support
• Invest in universal math screenings and evidence-based interventions
• Support research and development to advance math instruction

Although math proficiency rates on the United States’ National Assessment of Educational Progress (NAEP) have increased since 2022, they remain below the pre-pandemic rates of 2019 for fourth and eighth grades. In math, from 2009 to 2019, student performance on the NAEP was stagnant, with roughly only 40 percent of fourth graders, 34 percent of eighth graders, and 25 percent of 12th graders at or above the proficient level.⁹ Following the COVID-19 pandemic, NAEP proficiency rates dropped for all three grades,¹⁰ further igniting concerns about student math achievement and discussions on how to improve math instruction.¹¹ NAEP defines proficiency as solid academic performance consisting of demonstrated “competency over challenging subject matter including subject-matter knowledge, application of such knowledge to real world situations, and analytical skills appropriate to the subject matter.”¹²  On the most recent NAEP, administered in 2024, only 39 percent of fourth graders scored at or above the proficient level¹³; performance was even worse for middle grades, with just 28 percent of eighth graders demonstrating math proficiency.¹⁴ By the time they leave high school, only 22 percent of 12th graders have fluent math skills.¹⁵

When these proficiency rates are broken down by socioeconomic status, the results are even more concerning: Only 25 percent of economically disadvantaged fourth graders,¹⁶ 14 percent of economically disadvantaged eighth graders,¹⁷ and 10 percent of economically disadvantaged 12th graders performed at or above proficient level.¹⁸

But students are not just performing poorly when measured by U.S. standards. To get a global perspective of student performance, the OECD’s PISA assesses 15-year-olds around the world in reading, math, and science every three years.¹⁹ In results on the PISA administered in 2025, the United States for the first time in the past 15 years achieved the exact same score as the OECD average in math—largely due to declining performance of other developed countries given the U.S average math score was the country’s lowest on record—and ranked 27th out of the 90 participating education systems.²⁰ Over the past 15 years, the United States has consistently scored below average in math while countries including China, Singapore, Japan, and Korea have ranked at the top and countries such as Canada and Estonia have routinely scored above the OECD average. Meanwhile, U.S. rankings over the past 15 years are:

• 2009: ranked 31st out of 65²¹
• 2012: ranked 36th out of 65²²
• 2015: ranked 40th out of 70²³
• 2018: ranked 37th out of 78²⁴
• 2022: ranked 34th out of 81²⁵
• 2025: ranked 27th out of 90²⁶

Rebounding from its record-low math performance will demand years of sustained effort from the United States, particularly as American students are estimated to lag more than five years behind their counterparts in China.²⁷

Several systemic factors have contributed to the decline in U.S. math performance, including challenges with hiring well-qualified math teachers, particularly in lower-income and rural areas;³⁷ a shortage of universal math screening assessments compared with reading;³⁸ inconsistent standards in teacher preparation programs; and limited access to high-quality instructional materials and implementation support for teachers.³⁹ Lowering proficiency cutoff scores on state assessments has further obscured this problem, creating an honesty gap that has masked falling achievement.⁴⁰ These shortcomings leave students at an academic deficit and unprepared for higher education and the workforce. Among students designated as college-ready by their high school coursework, only 47.6 percent are projected to pass intermediate college algebra.⁴¹ Consequently, colleges nationwide, including elite institutions such as Harvard University, have had to establish introductory and remedial math courses to close foundational algebraic skill gaps.⁴²

Although early math skills such as measurement and pattern recognition in kindergarten strongly predict eighth grade math outcomes, science and reading achievement, and grade retention, access to advanced coursework in middle and high schools remains limited.⁴³ During the 2020-2021 school year, 26 percent of students taking Algebra 1 were in middle school while 73 percent were in high school, despite Algebra 1 being a prerequisite for advanced level math and science courses.⁴⁴ Furthermore, just 48 percent of U.S. high schools offer calculus, a figure that drops to 35 percent in schools with high enrollment of Black and Latino students.⁴⁵

Addressing these deficits has significant long-term employment and economic benefits. Improving math performance by 0.5 standard deviations before age 12 correlates with higher earnings by age 30 across all demographic groups, with girls and Hispanic students seeing the largest relative gains.⁴⁶

Boosting mathematics achievement across the United States is critical to student success,⁵⁰ economic mobility,⁵¹ national security,⁵² and global competitiveness,⁵³ especially given the rapid growth of technology and the development of new jobs requiring technology-related skills that rely on math proficiency.⁵⁴ To prepare students for the future, federal and state policymakers must work with district leaders to advance and implement evidence-based practices that strengthen math instruction at all grade levels. Below are five recommendations that policymakers and state and school district leaders should consider.

Measuring how well students are performing in math and holding stakeholders accountable is essential to increasing math achievement. Policymakers should strengthen accountability, academic standards, and transparency in K-12 math by enhancing statewide accountability systems, bridging secondary and postsecondary math expectations, and modernizing math pathways and advanced course offerings.

Policymakers must ensure that statewide accountability systems integrate comprehensive metrics that measure student growth and proficiency in math and other factors that affect math success, including educator effectiveness and equitable access to advanced math coursework. Accountability frameworks should also track long-term postsecondary outcomes, connecting K-12 math achievement to college readiness and employment outcomes.

Strengthening academic standards in math, starting in early childhood, should be a priority for state and local policymakers. In 2018, Mississippi developed early learning standards to include mathematics for classrooms serving infants through pre-K students, while Alabama has historically integrated mathematical thinking standards into its early childhood framework with evidence that students enrolled in its pre-K programs demonstrate higher proficiency in reading and math and statistically significant academic benefits into elementary and middle school.⁵⁵ This suggests that building math fluency skills must start early, well before statewide testing begins in third grade. Academic benchmarks and standards across early childhood and elementary and middle schools should be aligned to ensure that all students are prepared to complete Algebra 1 by eighth grade. States should also consider adopting auto-enrollment policies for math, which allow schools to automatically enroll any student who meets the criteria for advanced level courses based on various measures including performance in preceding advanced courses and results on statewide standardized exams. These policies ensure higher-achieving students are enrolled in advanced math coursework without needing a recommendation from a counselor or other professional, eliminating delay and reducing gender and equity gaps.⁵⁶ This will require modernization of academic standards and curricula to include courses in high-demand modern disciplines such as applied statistics, data science, and probability, which are tailored to careers in technology, specifically artificial intelligence (AI) and engineering. By developing clear advanced math pathways and identifying early course sequences for advanced math, states will ensure that students are prepared for post-calculus coursework before they leave high school.

State leaders should also strive to align secondary and postsecondary expectations by connecting high school graduation standards with NAEP 12th grade math proficiency benchmarks, content knowledge and mathematical reasoning skills of the PISA, and entry-level college math prerequisites to help students avoid remedial courses during their first year of college. Aligning expectations should also include increasing the number of math credits or courses that are required for high school graduation. As of 2023, only 17 states and Washington, D.C., require four years of math credits or courses, while the majority of states require only three, and a few states, such as California, require just two.⁵⁷ State policymakers should consider mandating at least four years of math for graduating seniors.

Access to rigorous, standards-aligned curricula across classroom settings and ongoing training and support for teachers are essential to ensuring instructional coherence and the connection of all aspects of a student’s learning, including standards, curricula, instructional practices, and assessment.⁵⁸

Policymakers should invest in initiatives that support districts in properly vetting and adopting standards-aligned curricula across grade levels and subjects. This may include investing in technical assistance personnel to advise districts and schools in the adoption and implementation of new curricula. Additionally, districts should provide individual schools with instructional math coaches. Coaches can provide targeted support to teachers to help with the rollout of new curricula and ensure daily use of high-impact instructional practices such as problem-based learning that engages students in deep thinking⁵⁹ and differentiated instruction, which has been found to help increase students’ engagement and math comprehension.⁶⁰ This use of instructional math coaches has been found to have positive effects on both student achievement and teacher confidence.⁶¹

To address poor math achievement in the United States, math instruction must be not only rigorous but also engaging and relevant to students.⁶² Students commonly report being uninterested in math class and finding math instruction to be disconnected from real-life skills, limiting their engagement in learning.⁶³ Integrating technology and hands-on learning into instruction will not only increase student engagement, but also improve math comprehension and prepare students for the future workforce.

Research has shown that manipulatives,⁶⁴ objects that can be used as teaching tools to understand and explore mathematical concepts,⁶⁵ can increase both student engagement and achievement when appropriately used in instruction.⁶⁶ Manipulatives include objects such as block sets and fraction tiles—anything a student can touch and manipulate to visualize and connect mathematical concepts. These tools help students both conceptualize math and understand the appropriate vocabulary to explain mathematical ideas, a critical component of math achievement.⁶⁷ DeKalb County, Alabama, saw just how impactful the use of manipulatives can be when it reenvisioned math instruction during the COVID-19 pandemic.⁶⁸ The county saw math achievement dramatically increase while the majority of the country saw scores plummet.⁶⁹ The district used pandemic relief funding to purchase and implement manipulatives in classroom instruction, as opposed to a traditional reliance on pencils and worksheets, and provided teachers with coaching on instructional strategies and data collection and analysis⁷⁰. These practices, which make math more “real” to students and foster critical thinking, are also common in Singapore and other Asian countries with high-performing school systems.⁷¹ In Singapore, students are encouraged to work in pairs or small groups to use mathematical skills in solving real-world problems and to engage in questioning and dialogue to support their understanding of math concepts.⁷² State and federal policymakers can support districts in making this instructional transition by providing designated funding to purchase manipulatives and training resources that guide school leaders and teachers on the most effective uses of manipulatives, including when and how to use different types.

Similar to hands-on and project-based learning, embedding standard-aligned technology into math instruction can increase engagement,⁷³ strengthen student learning,⁷⁴ and give students the opportunity to learn about the intersection between math and technology. Louisiana, where fourth graders ranked second for growth in math achievement in 2024,⁷⁵ launched a statewide math initiative in 2023 known as “Math Refresh.”⁷⁶ This initiative included free access to Zearn Math for all K-8 teachers.⁷⁷ Zearn Math is a digital tool used to supplement math instruction through guided lessons, games, and assessments.⁷⁸ Initial studies of the partnership in Louisiana found that students using the platform scored nearly six points higher on the state math assessment than students who did not use it.⁷⁹ Other schools are choosing to adopt new innovative technologies⁸⁰ such as Prisms, a virtual reality tool that presents students with real-life problems and guides them through mathematical exercises to find a solution.⁸¹ While larger studies are currently underway, ⁸²one study has found that students who completed lessons using Prisms scored 11 percent higher on a researcher-designed assessment than those who did not, and their teachers reported higher levels of student engagement.⁸³ State investments in evidence-based, standards-aligned technology can assist schools in finding and implementing effective tools that increase student engagement and learning. When adopting new technologies, states should be sure to include various stakeholders, including educators, school leaders, and parents, in the evaluation and decision-making processes. States should also provide additional funding to Title I of the Elementary and Secondary Education Act (ESEA) and to rural schools that may not have available funds to purchase the necessary equipment.

Between 2012 and 2020, the U.S. supply of qualified teachers dropped, with the number of graduates of teacher preparation programs prepared to teach math declining by 36 percent.⁸⁴ Moreover, only 1 in 8 preparation programs currently prepares prospective elementary teachers for effective math instruction.⁸⁵ Most programs provide an average of 85 instructional hours in specialized mathematics content knowledge and 51 hours of math pedagogy,⁸⁶ 14 hours total below the recommended 150 hours outlined by experts.⁸⁷ This lack of preparation has a real impact. Research finds that students perform worse when their teacher is less confident in their ability to deliver math instruction effectively.⁸⁸ To address these challenges, lawmakers should incentivize the math teacher pipeline, standardize preservice math curricula, and implement strategies to retain effective math educators.

The inadequate preparation and shortage of qualified math teachers require expanding federal and state programs that offer tuition assistance and grants to incentivize more young people to pursue degrees in math education. Leaders should also seek to expand accelerated certification pathways to attract STEM professionals and tech workers displaced by the AI boom to transition into teaching. These pathways should offer flexibility for endorsements and licensure for STEM and technology professionals who already have knowledge and competencies in mathematics, as well as competitive compensation structures and support to effectively teach.

The math content areas and teaching methods taught in teacher preparation programs are also critical to student success. Only 21 states currently offer detailed guidance to preparation programs on what they should teach in specific math content areas.⁸⁹ State lawmakers and licensure and certification commissions should establish uniform coursework, clinical experience standards, and licensure exams aligned with the National Council of Teachers of Mathematics core domains in number and operations, algebra, geometry, measurement, and data analysis and probability.⁹⁰

Leaders should also implement mechanisms that increase pay for effective teachers and provide professional learning communities to support math educators and promote their retention.⁹¹ These structures are needed for all teachers, but most importantly for elementary teachers who do not feel prepared to deliver strong math instruction and lack specialization in mathematics,⁹² as well as early-career teachers who could benefit from creative classroom staffing that can provide additional support during their first years of teaching.⁹³

Before students can receive interventions, they first must be identified for math deficiencies. The earlier students are identified and given targeted support, the better.⁹⁴ To improve math achievement, federal and state investments first must be directed toward universal screenings.⁹⁵ States including Iowa, Alabama, and Indiana have all passed laws mandating that students in early grades, often kindergarten through second or third, are screened regularly to identify and address academic deficiencies in math.⁹⁶ Federal investments in the form of grants should be provided to districts to acquire research-based universal screening tools, improve data collection systems, and provide related training to educators and staff. Screening tools should include both computer-adaptive tests and curriculum-based measures⁹⁷ and be available in multiple languages to meet the needs of individual students. These tools should also be aligned with state standards and assessments to accurately predict how students will perform on end-of-year statewide assessments and to provide targeted interventions.

After students have been screened, districts should be required to use these data to deliver evidence-based interventions such as high-dosage tutoring and other academic interventions.⁹⁸ Early identification of students who are behind allows them to access adequate interventions, which are essential to reversing the decline in student math achievement.⁹⁹

Research consistently shows that high-dosage tutoring is one of the most effective interventions for both reading and math.¹⁰⁰ For example, high-dosage tutoring in Chicago Public Schools in Illinois and Fulton County Schools in Georgia has been found to undo pandemic math learning loss, with math tutoring having a total impact equivalent to two-thirds of a year of learning (p=0.03).¹⁰¹ In Massachusetts, where students have historically achieved among the highest scores on national math assessments,¹⁰² state leaders have also acknowledged high-dosage tutoring as a piece in their formula for success.¹⁰³ This style of tutoring differs from traditional tutoring in that it requires small groups meeting at least three times per week during school hours with a consistent, professionally trained tutor using high-quality materials that are aligned with classroom content.¹⁰⁴ However, establishing a high-dosage tutoring program and vetting legitimate providers can be challenging for schools and districts to take on by themselves.¹⁰⁵ To address these barriers, Massachusetts has approved math tutoring providers for districts to choose from and provided competitive grant funding for which districts can apply to cover the costs.¹⁰⁶ Other states—such as Tennessee, where fourth grade students have bounced back to pre-pandemic math achievement levels¹⁰⁷—have developed statewide high-dosage tutoring programs to expand interventions.¹⁰⁸ The use of high-dosage tutoring has surged over the past five years in response to pandemic-related learning loss,¹⁰⁹ fueled by historic levels of federal relief funding.¹¹⁰ However, this funding was abruptly rescinded in the spring of 2025, leaving states scrambling to find funds to cover the costs of programming.¹¹¹ To ensure all students continue to have access to high-dosage tutoring, the federal government should strengthen federal programs, such as Title I of ESEA and the Student Support and Academic Enrichment grant program, to fill gaps in funding and prevent disruptions in service.¹¹²

Extended learning time is another strong intervention for students falling behind. Among the many evidence-based investments of the Alabama Numeracy Act is that it requires each local education agency to provide a summer math camp at no cost for elementary students identified as having a math deficiency,¹¹³ a strategy that increases instructional time and has been found to have a positive effect on math achievement.¹¹⁴ For early grades, the program embeds math instruction into literacy camps.¹¹⁵ For fourth and fifth graders, however, the program requires a separate math camp that delivers between 40 and 70 hours of instruction from a highly effective educator.¹¹⁶ Schools with the lowest math performance receive state funding to cover staff and operations.¹¹⁷ While the program is promising, the state’s latest annual report found that just 17 percent of eligible fourth and fifth graders in the state’s full-support schools attended a summer math camp in 2025.¹¹⁸ States looking to replicate this programming should encourage participation through informational sessions and offer resources that may reduce attendance barriers, such as transportation and convenient scheduling for working families.

Math education must prepare students for an AI-driven economy. Therefore, federal investments should support interdisciplinary research that connects K-12 education practice to AI, data science, and technological advancements. Grantmaking at the National Science Foundation has already started to forge collaboration across these disciplines but more initiatives within other federal programs are needed.¹¹⁹

Although states are rapidly adopting the science of reading and requiring universal literacy screeners,¹²⁰ few states are requiring universal numeracy screeners and the debate continues on the best way to teach math.¹²¹ Universal numeracy screeners are instruments used to detect learning gaps and identify students at risk for math failure.¹²² These tools serve as an early intervention catalyst to provide students with support early on when it is most effective.¹²³ To strengthen math instruction in the United States and reverse the decline in achievement, the federal government should invest in research and evaluation to support states in identifying and adopting evidence-based numeracy screeners and evidence-based instructional practices.

Research on math instruction is lagging when compared with reading instruction.¹²⁴ Without scientific evidence on how to effectively teach math, declining achievement cannot be reimagined and revamped across the nation. Federal lawmakers must provide adequate funding for research on effective math instruction from early childhood through higher education. This research should identify how students best learn math, taking into account not only the recommendations discussed above but also cognitive psychology and the science of learning.¹²⁵ This body of knowledge, known as the science of math,¹²⁶ should be disseminated to states and districts in the form of guides and practical tools that can be implemented easily at the local level to address the common gap between research and practice.¹²⁷ Currently, however, the U.S. Department of Education’s research branch, the Institute of Education Sciences (IES), is limited in its ability to rapidly produce the scale of research needed.¹²⁸ To bolster IES and give it the tools it needs to effectively and rapidly conduct the necessary research to strengthen math education, Congress should consider legislation that increases IES’ capacity. For example, the New Essential Education Discoveries Act would establish a new center under IES,¹²⁹ similar to the Defense Advanced Research Projects Agency, to tackle education’s biggest issues, including math instruction, through informed-risk, high-reward research projects.¹³⁰ A center such as this could lead the necessary research to advance math instruction rooted in the science of learning. The act would also upgrade education data systems to better track student progress and offer high-quality data that researchers can use, an important aspect of effective research and development.¹³¹ With this research, schools can strengthen math instruction using evidence-based strategies to make gains similar to those seen in literacy in Mississippi, Louisiana, and other states that have adopted the science of reading.¹³²

Math education has suffered historic underinvestment compared with reading education. Although the science of reading has expanded at both the state and federal levels¹³³, the distribution of resources for early screening and evidence-based interventions in math to schools and districts has lagged. As AI transforms the modern workforce,¹³⁴ policymakers must urgently prioritize math education to equip U.S. students with the skills needed to compete globally against high-performing students from countries such as China, Singapore, and Japan.¹³⁵ At the federal level, lawmakers should expand funding for teacher preparation programs to attract top-tier math educators and incentivize STEM professionals displaced by AI to transition into the classroom. Rather than weakening accountability,¹³⁶ federal policymakers must strengthen oversight across the school, district, and state levels by incorporating comprehensive math metrics that guarantee access to high-quality math instruction that leads to improved postsecondary outcomes. Congress must also reinvest in education research and development, particularly by passing legislation to bolster IES. Strengthening IES would expand the science of math and bridge the gap between research and practice.¹³⁷

However, federal action alone cannot reverse declining math achievement. State and local leadership—governors, legislatures, and education boards—must play a role. States such as Massachusetts and Alabama offer solid blueprints for increasing math performance that other states can adopt and scale. With federal funding and technical assistance, states and districts should adopt universal numeracy screeners and high-quality instructional materials, ensuring instructional coherence by aligning academic standards, curricula, instructional practices, and assessments. State licensure boards and teacher preparation programs should establish rigorous instructional-hour requirements that are aligned with the National Council of Teachers of Mathematics and other subject-matter experts.¹³⁸ States should also expand math pathways by enacting auto-enrollment policies for advanced math tracks and offer modern math courses that are tied to fast-growing fields such as data science and engineering. Finally, policymakers should raise graduation standards and require a minimum of four years of math courses for high school graduation and completion of Algebra 1 by eighth grade. Given the rise of AI, investing in math education is no longer optional; it is a direct imperative for long-term economic competitiveness.

The authors would like to thank Mishka Espey of the Center for American Progress for their valuable contributions to this report and Madison Weiss for her thorough fact-checking.