This study provides evidence on the relative effectiveness of computer-assisted learning (CAL) software and traditional teaching. Based on a field experiment in Salvadoran primary schools, we evaluate three interventions that aim to improve learning in mathematics: (i) additional teacher-led classes, (ii) additional CAL classes monitored by a supervisor, and (iii) additional CAL classes instructed by a teacher. We find that CAL lessons lead to larger learning gains and are less sensitive to class size as well as student ability than teacher-centered classes. Our results highlight the value of CAL in an environment with heterogeneous classes and poorly qualified teachers.