This paper investigates the influence of the school shift change on the academic performance of 6th and 7th graders (lower secondary education) of public schools in the city of Recife, Brazil. The empirical analyses use educational data from a panel of the Brazilian Ministry of Education, with a sample of 4,500 students, 3,468 parents or guardians, 85 principals, and 137 teachers working in 87 public schools that were spatially distributed. The identification strategy explores a quasi-experiment in which some classes of the investigated sample had their school shifts changed between the two years, which occurred exclusively due to the school logistics and regardless of the parents’ preferences. Difference-in-differences models combined with propensity score matching demonstrate that students who shifted from morning to afternoon between the two years analyzed improved their Portuguese test scores, and that those who experienced the opposite change, from afternoon to morning, had a drop in the performance. Estimations in intermediary outcomes also reveal that the changes in the students’ sleeping habits are an important mechanism that explains the relation between the scores and the school shift.