A peer-reviewed study of Khan Academy has been published in Proceedings of the National Academy of Sciences (PNAS), one of the most selective scientific journals in the world. Using a methodology designed to account for factors that are typically not measured in education research, the study finds a statistically significant positive relationship between Khan Academy usage and student learning outcomes.
For school leaders evaluating edtech tools used at scale, PNAS sets an unusually high bar: it does not publish routine program evaluations or vendor-led efficacy studies. Submissions are reviewed under criteria more commonly applied in fields like medicine or economics, often by members of the National Academy of Sciences—a distinction that includes many of the world’s most highly cited scientists and Nobel laureates.
For school leaders evaluating edtech tools used at scale, PNAS sets an unusually high bar: it does not publish routine program evaluations or vendor-led efficacy studies.
The analysis draws on multiyear data from roughly 200,000 students across hundreds of U.S. school districts and captures how Khan Academy was used across a wide range of real classroom contexts. In this sample, average use amounted to about 10–15 minutes per week, which is below the 30 minutes-per-week threshold many educators aim for and what we recommend.
Rather than relying on simple usage comparisons, the analysis explicitly accounts for many factors that are often unmeasured in education research, including student motivation, prior achievement, tutoring access, teacher practices, and year-to-year changes in instructional context. The methodology was developed in collaboration with researchers from Stanford University and the University of Toronto. Our head of efficacy research, Bogdan Yamkovenko, shares a detailed discussion of the methodology and modeling choices in a separate post.
After accounting for all of these sources of noise in the data, increased weekly usage of Khan Academy is associated with higher MAP Growth gains, a widely used measure of student learning.
If you’ve reviewed edtech research before, you’ve likely learned to be cautious—many studies rely on small samples, short pilots, or vendor-led analysis. In practice, school district leaders tend to look for independent, national-scale studies with conservative methods. That’s the bar this research was designed to meet.
In practice, school district leaders tend to look for independent, national-scale studies with conservative methods. That’s the bar this research was designed to meet.
If a learning tool is going to be used by millions of students, it should be evaluated accordingly—with transparent methods, conservative assumptions, large and representative samples, and independent peer review.
Taken together, these findings provide national-scale evidence that sustained use of Khan Academy can support measurable learning gains in real classroom conditions. For districts making high-stakes adoption decisions, this study offers an unusually rigorous, independent signal of impact—and points toward a higher standard for evaluating tools used at scale.
