Online tutoring reduces by half the learning loss due to school closures: Evidence from a randomized experiment in Kenya

We evaluate the effects of an online tutoring program that started in 2016 and continued during the pandemic despite the schools being closed for 9 months in Kenya. Using videoconferences, volunteer students from a Canadian university tutored grade 6 students (12 years old) in a rural school in Kenya, on the topics of Maths and English. We implement a randomized experiment to test the effects. We find no effect when the schools are open, but a large effect when the schools are closed (0.4 SD increase in exam scores in the treatment group versus control group). Since we have data from before the pandemic, we are able to quantify the learning loss due to COVID-19: 0.8 SD. We conclude that online tutoring compensates half of the learning loss. • We evaluate the effects of an online tutoring program in Kenya. • We find no effect on tests scores when the schools are open, but a large effect when the schools are closed. • Online tutoring compensates half of the learning loss due to school closures during COVID.