Concerns about methodology used in real world experiments on transport and transport safety

Abstract

Real-world experiments in which some experimental units (e.g., people or roadway sites or vehicles) are randomized to treatment and others to no treatment are increasingly regarded as practicable and desirable. This is the case for complex educational and social welfare interventions as well as for therapeutic drug trials. The present paper discusses the relevance of such experiments to transport and transport safety research. When compared with the more common designs of research, both the advantages and the disadvantages may seem substantial, and decisions will require good judgment. Drawing lessons from medical and social welfare research, the paper aims to be a resource when the relative merits of randomized experimentation and other methods are being debated.