Panorama des méthodes associées à l'évaluation de type "avant- après" = Overview of methods related to "before-and-after" evaluation.

Scientific evaluation of a traffic safety measure applied to a set of road systems or road users is based on three statis- tical processes (1) survey design (2) experimental design, and (3) statistical inference. All statistical procedures currently in use are reviewed, with particular attention on paradigmatic analysis of before and after data from controlled experiments.