Estimating the effect of remote teaching for university students through generalised linear mixed models

The present paper aims at analysing the effects of remote teaching on university students’ careers, as a consequence of university closures due to the COVID-19 pandemic. For this purpose, we use administrative data of the University of Florence on students’ careers and compare their performance in terms of the probability of passing specific exams. In particular, using a random intercept logit model, we compare the group of students enrolled in the academic year 2018/2019 – who received classic face-to-face teaching – with the group of students enrolled in the subsequent academic year, who experimented remote teaching during the second semester. Results obtained on different degree programs show that the effect of remote teaching at the course level is markedly heterogeneous, with different sign and magnitude.