Accounting for repetition and dropout in contemporaneous cross-section learning profiles: Evidence from Rwanda

•Learning profiles allow us to trace learning progress through schooling.•In this paper I estimate a learning profile for Rwanda using a household-based survey of children.•I use a detailed schooling history to show that observational learning profiles that don’t explicitly account for repetition and dropout may substantially over-state learning gains per year. How much do children learn in a year of school? Longitudinal data that tracks children over time is scarce in developing countries, and so recent studies estimate learning profiles by comparing the ability of people with different amounts of schooling, at a single point in time. Such estimates of the effect of schooling on learning may be biased upwards by not controlling for repetition and dropout. In this paper I estimate contemporaneous cross-section learning profiles for Rwanda, using data from a nationally representative survey of 3053 children aged six to eighteen. I show how adjusting this learning profile for the total number of years enrolled in school (accounting for repetition and dropout), using detailed schooling histories, reduces the average amount learnt per year by over 60 percent. The learning profile for Rwanda is not just too flat, but flatter than previous estimates suggest.