Using a Field Experiment to Understand Skill Formation During Adolescence

Abstract

We combine a field experiment with structural estimation to study skill formation during adolescence. We randomize nearly 1,000 adolescents from low-income communities to distinct low-cost classroom interventions that lasted 10-11 weeks. We find that investments during adolescence can improve skills, particularly executive functioning skills. We estimate production functions for each treatment group to uncover the underlying mechanisms. The technology of skill formation itself is impacted by our interventions. Improvements in cognitive skills were driven by increases in returns to scale and the relative importance of investments, while improvements in executive functioning skills were driven by a level increase in efficiency. We find important heterogeneity in these mechanisms by initial skill levels, and illustrate that predictions about program impacts on high school graduation depend crucially on (i) the persistence of the change in technology (ii) how early the program starts, and (iii) the weight on executive functioning relative to cognitive skills in determining high school graduation. Our results suggest that the policymaker not paying attention to the adolescent years will miss out on a key sensitive period.

Keywords: field experiment, technology of skill formation, cognitive ability, executive functions.

JEL Classification: I24, I28, J13, J24, C93.