Returning data is not enough: Feedback, managerial beliefs and student outcomes

Abstract

This article assesses whether a very low-cost, purely informational diagnostic feedback intervention can improve learning and internal efficiency in primary schools in a middle-income country. We ran a randomised controlled trial in 712 schools in Mendoza (Argentina), where headteachers in the treatment group received comparative reports on enrolment, promotion, repetition and dropout, built from administrative records, while control schools accessed the same information one year later. We combined these reports with standardised tests in Mathematics, Language, Natural Sciences and Social Sciences, and with surveys of headteachers before and after the intervention, estimating intent-to-treat (ITT) effects with stratification and school-clustered errors. We find no significant impacts on promotion, repetition or dropout, and small negative effects of around 0.05 standard deviations in Mathematics and Language. Headteachers also hold systematically biased perceptions of internal efficiency indicators, which only weakly converge towards the administrative data. The findings provide experimental evidence on the limits of interventions based solely on returning information, and suggest that information without complementary management support is unlikely to generate substantive short-term improvements in academic outcomes.