The aim of this translational research was to investigate the effect of differential training in three types of rule-governed behavior (RGB) on compliance and transgression with pedestrian traffic rules. Participated 231 university students divided in three groups defined by the type of training (pliance, tracking and augmenting). The design consisted of pre-tests and post-tests, in addition of a training phase with direct reinforcement according to the experimental group. The results showed a general trend by compliance the rule in all phases, particularly with tracking-consequences in pre and post-training. In all experimental groups, training in RGB increase election in compliance of traffic rules. These findings allow the formulation of evidence-based guidelines for the design, implementation and evaluation of prevention road programs.