Can Tutoring—and Technology—Finally Solve Bloom’s Two Sigma Problem?

By: Harry Anthony Patrinos, Lelys Dinarte, James Gresham, Renata Freitas Lemos, Rony Rodrigo Maximiliano Rodriguez Ramirez

In education research, few findings have been as provocative—or as frustrating—as Benjamin Bloom’s two sigma problem. In the 1980s, Benjamin Bloom and his students demonstrated how learning could be transformed. He found a method for delivering learning outcomes that are improved by a factor of two standard deviations (two sigma). This would come about by combining two approaches: mastery learning and one-to-one tutoring. Mastery learning means that each student must achieve true proficiency in a topic before moving on to the next, more advanced subject, even if that takes longer than other students. One-on-one tutoring means that each student is provided with a personal tutor who guides them through their learning.

The catch is that scaling tutoring for every student seemed impossible. For decades, reformers tried other strategies—smaller classes, different curricula. None came close.

Today, we may finally have a breakthrough. Technology is making tutoring affordable, scalable, and—most importantly—effective.

Tutoring Works Better Than Almost Anything Else

Tutoring has been around a long time. In antiquity and the Middle Ages this is how the elites were educated, hiring instructors privately for individuals or small groups. In recent times it has become synonymous with test preparation, or cram schools, and fallen out of favor. But what we are discussing here are attempts to raise the achievement of students who have fallen behind. Therefore, tutoring is defined here as one-on-one or small-group instruction by teachers, paraprofessionals, volunteers, or parents.

In the United States, tutoring programs yield consistent and substantial positive impacts on learning outcomes, with large effect sizes up to 0.55 standard deviations. That is much more than a year’s worth of learning. The fraction of statistically positive treatments is larger than early childhood interventions.

But the details matter. High-dosage tutoring—four or more days a week, groups of six or fewer—produces the strongest results. It works because students get more time, more personal attention, and better-aligned instruction.

The problem is cost. In-person tutoring can run $1,800–$4,000 per child each year. A nationwide rollout could top $15 billion.

As I have written before, given how high the returns to education are, the real question is not whether to invest—but where and how much.

Online Tutoring: Scaling the Gold Standard

When the COVID-19 pandemic hit in 2020, schools everywhere scrambled to find ways to keep students learning. Tutoring went online almost overnight. What’s remarkable is not just that it worked, but that it worked almost as well as in-person tutoring—but at a fraction of the cost.

Recent studies make the case.

Take Italy. During lockdowns, university students volunteered to tutor disadvantaged middle schoolers online, three to six hours a week. A new study released this week shows striking results. Math scores rose by 0.22 standard deviations in 2020 and 0.20 in 2022. Students who received more tutoring hours gained the most, while group formats diluted the effect. But the story went beyond test scores. Aspirations and socioemotional well-being improved—showing that a screen connection could also provide hope and human contact in the middle of isolation.

If Italy demonstrated the power of online tutoring in a health crisis, Ukraine has shown that it can even work in the middle of a war. Between early 2023 and mid-2024, three randomized experiments reached nearly 10,000 Ukrainian students across all regions of the country. For six weeks, children in grades 5-10 received three hours per week of small-group tutoring in math and Ukrainian language. Despite the chaos of displacement, blackouts, and air raid sirens, the program delivered learning gains of 0.49 SD in math and 0.40 SD in language, along with meaningful reductions in stress. Parents and students embraced it. Teachers saw students re-engage. And because it relied on existing infrastructure and teachers as tutors, the program was highly cost-effective.

In other words, online tutoring didn’t just help during school closures. It worked during wartime. It gave children stability, learning, and a reason to believe in the future. That’s an extraordinary achievement, and it suggests online tutoring has resilience and adaptability that few other interventions can match.

Spain offers another example. The program paired middle schoolers with qualified teachers for two-to-one online tutoring, three hours a week for eight weeks. Pupils improved both in math and in socio-emotional well-being.

The lesson from Italy, Ukraine, and Spain is the same: online tutoring can deliver the benefits of in-person tutoring at scale.

[While the OECD and Ukraine examples show that online tutoring can scale in advanced and middle-income contexts, there is also evidence from low-tech interventions. In Bangladesh, phone-based mentoring raised learning by 0.75 SD for just $20 per child. In Botswana, SMS and phone tutoring produced the equivalent of a full year of schooling for $100, with similar results replicated in five other countries. Sometimes, the smartest technology is the simplest one.]

Blended Learning: The Next Frontier

Tutoring is not the only piece of the puzzle. Many schools face a chronic shortage of qualified teachers, especially in rural areas. Blended learning—mixing technology with traditional teaching—may be the best way to fill the gap.

Studies in underserved areas show that hybrid models (broadcast lessons, pre-recorded videos, TV-based instruction) significantly improve outcomes.

China’s Dual-Teacher program connected rural classrooms with elite urban teachers via online lessons. Over three years, students gained nearly a full standard deviation in math—at low cost.

Blending learning also allows us to implement proven pedagogical approaches, whether they are called teach at the right level, response to intervention, personalized learning, direct instruction. Take a look at what Ørestad Gymnasium is doing in Denmark, or High Tech High in San Diego, or Apex High School in South Africa

Blended learning does not replace teachers; it amplifies them—bringing high-quality instruction to places where it would otherwise be absent.

What Policymakers Should Do

Bloom’s two sigma challenge is still with us. But we are closer than ever to solving it. The implications for policy are clear:

1. Invest in tutoring first. Few interventions produce such consistent, large effects.
2. Make it affordable. Online models can cut costs dramatically.
3. Use blended learning. Where teachers are scarce, technology can help fill the gap.
4. Focus on scale. Programs that rely on simple tools and existing infrastructure are the ones that last.

The irony is that we’ve poured billions into reforms that barely moved the needle, while tutoring—perhaps the most powerful lever we have—was neglected because of cost and logistics. That excuse is no longer valid.

Do we have the courage to scale online tutoring?

Lelys Dinarte-Diaz is a research economist in the Human Development Team of the World Bank’s Development Research Group. Her primary research fields are development economics and economics of education, with a focus on violence and crime. She uses experimental methods to study the impacts of psychology-based interventions on mental health outcomes. She has research projects in El Salvador, Guatemala, Honduras, Mexico, Jamaica, Peru, and Ecuador. Lelys obtained her Ph.D. and master’s in economics from Pontificia Universidad Católica de Chile and her B.A. in Economics from ESEN in El Salvador, the country where she was born and raised.

Renata Lemos is a Senior Economist in the World Bank Education Global Practice where she has worked on operations and analytical activities in Costa Rica, Ecuador, Mexico, Peru and Tanzania. Renata is a member of the core research team of the World Management Survey and her recent work focuses on topics in managerial and organizational economics in the public sector. Before joining the World Bank in 2016, she was a Lecturer in the Economics Department at Stanford University and a Research Associate at the Centre for Economic Performance, London School of Economics and at Harvard Business School. She holds a PhD in Land Economy (Applied Microeconomics) from the University of Cambridge.

James Gresham is an Education Specialist in the World Bank’s Education Global Practice. He has worked on operations and analytical and advisory activities in Europe, Central Asia, and Africa. His work focuses largely on strategies to improve learning outcomes, equity of service delivery, and the governance and efficiency of primary and secondary education. He has also worked on programs addressing school learning environments, early school leaving, and out-of-school youth. He is currently leading programs in Ukraine and Bosnia & Herzegovina, but has also worked in Uzbekistan, Romania, Moldova, Serbia, Kosovo, and Liberia. He is a doctoral candidate in Educational Administration and Policy Studies at the George Washington University’s Graduate School of Education and Human Development, and he holds an M.A. in International Affairs and B.S. in Business Economics.

Rony Rodriguez is a PhD student in Education Policy and Program Evaluation at Harvard University whose work centers on understanding and mitigating the impacts of major disruptions, such as pandemics, conflict, and displacement, on educational outcomes and human capital development. His research investigates how targeted educational and mental health interventions can strengthen student well-being and resilience, reduce inequality, and improve long-term educational and economic trajectories for vulnerable populations. He focuses in particular on school-based interventions like individualized tutoring, enhanced curricula, and integrated mental health and psychosocial support (MHPSS). He holds an M.D.P. in Development Policy from KDI School (South Korea) and a B.A. in Applied Economics from Universidad Centroamericana (Nicaragua).

Harry Anthony Patrinos is Head of the Department of Education Reform and 21st Century Endowed Chair in Education Policy at the University of Arkansas. He specializes in the economics of education, particularly the returns to schooling, school-based management, demand-side financing and public-private partnerships. He was formerly with the World Bank, where he was Senior Adviser, Education. He held management and technical positions at the World Bank. Previously, he worked in the Office of the Chief Economist for Europe and Central Asia. He managed education teams in Europe and Central Asia, East Asia and the Pacific, the Middle East and North Africa, and the Global Unit. He led lending operations and analytical work programs in Latin America. He co-led the development of the Harmonized Learning Outcomes database, part of the Human Capital Index, published in Nature. He has studied and worked extensively on the socioeconomic status of Indigenous Peoples. He has many publications in the academic and policy literature, with more than 50 journal articles. He previously worked as an economist at the Economic Council of Canada. Mr. Patrinos received a doctorate from the University of Sussex.

The post Can Tutoring—and Technology—Finally Solve Bloom’s Two Sigma Problem? appeared first on Getting Smart.