This study examines the scalability of A/B testing in educational settings, focusing on its application to improve the implementation and engagement of Mindspark, a computer-aided learning platform by Educational Initiatives in India. Mindspark, which serves grades 1–10 with personalized learning paths, has demonstrated efficacy through randomized controlled trials showing significant gains in Math (0.37 SD) over short periods and is currently being scaled in several states in India. However, challenges persist in sustaining teacher engagement and fidelity during large-scale implementation. Using a sample of over 20,000 students across 170 schools, we conducted iterative A/B tests to address critical implementation issues. Test rounds included report card-based interventions and phone-based calls to action, randomized at the school and teacher levels. Our results highlight A/B testing as a resilient and cost-effective method for optimising implementation at scale.
Meeting ID: 368 849 075 232
Passcode: xX3p6Ar2
