The Consequences of Promoting Data Literacy among Graduate Students

We study the impact of a data literacy program on graduate students' skills and academic outcomes at a large Italian university. The program targets students expected to have weak quantitative skills and provides 120 h of training in data interpretation and processing alongside the regular courses of their master's degree. Admission is rationed through random assignment of available slots among applicants. Exploiting this exogenous variation, we find that the program substantially improves data literacy among students with low initial numeracy. Despite the additional workload, participants do not experience slower academic progress in their main degree program and pass at least as many exams as students in the control group. We also find positive effects on GPA among students with lower initial numeracy. Finally, follow-up survey data collected about two years later show that low-numeracy participants are more likely to be in paid employment.