Despite the potential benefits of cooperative learning at university, its implementation is challenging. Here, we propose a theory-based 90-min intervention with 185 first-year psychology students in the challenging domain of statistics, consisting of an exercise phase and an individual learning post-test. We compared three conditions that manipulated the exercise phase: individual work, cooperative dyadic instructions (structuring three basic components of cooperative learning: positive goal interdependence, individual responsibility and promotive interactions) and cooperative dyadic interactions (the three basic components with an additional cooperative nudge, namely explaining why and how to cooperate in this task) in order to test whether a progressive increase in benefits occurs as the cooperative structure is reinforced. Results indicated a linear trend in individual post-test learning and competence perception, from individual work to cooperative instructions to cooperative interactions. Competence perception mediated the effect of experimental conditions on learning. The results highlight the benefits of the cooperative nudge.