Better than nothing: the effects of video-conferencing on learning as a substitute for in-person instruction

This paper aims to estimate the effects of remote classes on learning in a context where in-person classes are fully suspended, such as during the COVID-19 pandemic. Using data on the remote modalities each public school in Brazil implemented during the school closure period (2020-2021), and using a sample of schools that did not return to in-person classes during this period, I estimated average-treatment-on-the-treated effects on language and mathematics standardized tests using an event-study framework for having remote classes against not having classes, and for having video-conferencing classes against having only online recorded classes. Results indicate that only specific groups of the sample were affected by the treatments, with overall small effects: remote classes were more effective for schools in the Northeast region and for schools attended by low-achieving 5th-graders; video-conference classes outperformed online recorded classes in Mathematics in low-achieving fifth-grade schools, and in Portuguese in schools attended by students with higher socioeconomic status.