Abstract:
Learning outcomes remain low and stagnant in many low- and middle-income countries, despite high school enrolments. Although improving learning remains a persistent challenge for governments, growing evidence identifies several education reforms that can improve learning, such as targeting instruction to students' individual learning levels rather than teaching one-size-fits-all curricula. While evidence suggests high efficacy of targeting instruction in randomized controlled trials, less is known about how to achieve effectiveness when governments scale up these types of pedagogical approaches, where several constraints might bind. We optimize one crucial constraint: class size. Larger classes necessitate even more targeted instruction because they have greater heterogeneity, suggesting targeting instruction might yield larger marginal returns. Conversely, larger classes also make it harder to cater to all student levels due to this increased heterogeneity. We conduct two randomized A/B tests to optimize targeted instruction in South Africa testing various class sizes, ranging from 30 students per class to 45 students per class. We find larger class sizes have minimal effects on the effectiveness of targeted instruction, suggesting this constraint is not in fact binding, and that targeted instruction could scale to reach 50% more children at minimal marginal cost. These results inform a growing literature on the science of scaling effective programs, and also contribute novel evidence to a contentious and mixed class size literature.
 
About the presenter:
Noam Angrist is the Academic Director of the What Works Hub for Global Education at the University of Oxford. He is also the co-founder of Youth Impact, one of the largest NGOs dedicated to scaling-up health and education programmes backed by rigorous randomised trial evidence. The organisation has scaled evidence-based programmes to over 1 million youth across five countries and led the use of A/B testing in the social sector to optimise evidence-based programmes on the path to scale. Noam has published in leading academic journals including Nature, Science, the Review of Economics and Statistics, and the Journal of Economic Perspectives. He co-led the development of the World Bank Human Capital Index education pillar. He also led the academic research underpinning the inaugural report of the Global Education Evidence Advisory Panel, a global review recommending cost-effective 'smart buys' to improve learning outcomes in low- and middle-income countries. Noam has a B.S. in Mathematics and Economics from the Massachusetts Institute of Technology (MIT) and has a PhD (DPhil) from the University of Oxford.