lunch&LEARN: How Effective is Online Education for Gifted Students? Evidence from a Global RCT

In this session, Patrick Gaulé, Associate Professor in Economics at the University of Bristol, will go over a study in which he and his team evaluated the effectiveness of online education for mathematically gifted students using a randomized controlled trial with 620 participants.

Students were nominated by national Olympiad organization in 44 countries, representing the upper tail of global ability. They were assigned either free access to an 18-week combinatorics course valued at $600 from Art of Problem Solving (a leading platform that trains members of the U.S. International Math Olympiad team) or to independent study using an equivalent e-book.

His findings indicate that being assigned to the course increased final-exam performance by 0.15–0.19 SD (ITT). Using random assignment as an instrument for engagement, Gaulé and his team estimate that each month of participation increased scores by 0.20–0.26 SD among compliers—implying large gains of 0.9--1.2 SD for those completing the course.

However, persistence emerged as a key constraint: about two-thirds of assigned students engaged minimally, and observable characteristics poorly predicted engagement ($R^2 \approx 0.10$). These results suggest that online education offers a cost-effective way to reach gifted students globally, but scaling it up will require addressing the challenge of persistence.