After 40 years, digital math learning improves scores by just 0.2 standard deviations.

Schools spend heavily on educational technology, but past research shows confusing and conflicting results about whether it helps students learn mathematics. To make sense of this, researchers reviewed empirical studies published mainly between 2013 and 2026 across wealthy and low-resource settings. They examined how different tools and teaching methods actually work in classrooms. They found that technology gives a positive but modest boost to math achievement. How teachers use a tool matters more than the tool itself. Using devices for problem-solving and collaborative work helps students far more than simply replacing paper worksheets with digital screens. Furthermore, real learning benefits take time to appear. The largest, best-designed studies detect gains only in the second year of use, meaning shorter studies miss these improvements entirely. Finally, unguided access to generative artificial intelligencegenerative artificial intelligenceComputer systems that can create new content, like text or answers, that look like a human made them. can artificially boost immediate performance while hiding a drop in actual student learning.

Why it matters

For decades, schools have spent billions on digital math tools hoping to boost achievement. But the evidence shows only a small average gain, and it only appears after two years and with proper teacher support. Worse, unguided use of generative AI can actually hide learning loss, threatening to undermine basic skills.

Questions answered