Résumé
A number of studies, both in science education and in the humanities and social sciences, have shown that active learning has a more positive effect on student performance than direct instruction. As these studies were experimental or quasi-experimental, they compared ideal interventions designed and controlled by researchers. The present study aimed to measure the effects of active learning on student success under authentic conditions, by examining ordinary teaching practices that take place in real-life situations and are subject to various material, time-related and human constraints, which can affect teaching preparation and implementation. The study also investigated whether these effects depend on students' psycho-cognitive profiles. Data were collected from 356 teachers and 2168 students across various science and humanities programmes. The findings show that active learning under authentic conditions has a more positive effect on students' success in the first year of university than direct instruction. More specifically, they show that success is fostered when teachers adapt content to students' difficulties, use different active learning methods and in particular flipped classrooms, and use digital tools to engage students in activities. The results also show that active learning is equally beneficial for all students, regardless of their psycho-cognitive profile.