Evaluation of hazard perception skills among electric scooter riders in urban environments

The rapid proliferation of electric scooters (e-scooters) has transformed urban mobility worldwide. Despite becoming a prominent mode of urban transport, their rising use has been accompanied by increased crash involvement. Hazard perception, the ability to detect, evaluate, and anticipate impending traffic events, is a critical determinant of road safety but remains underexplored among e-scooter riders. This study examined hazard perception in novice, intermediate, and experienced riders using a video-based hazard detection task requiring button responses and verbal descriptions. Generalised linear mixed models analyses revealed that experienced riders were more sensitive to unmaterialised hazards than novices and intermediates, while no group differences emerged for materialised hazards. Novice riders also provided fewer verbal descriptions of unmaterialised hazards than both the intermediate and experienced riders, though no differences were observed for materialised hazard descriptions. Pathway type further modulated performance: materialised hazards were most likely identified on roads, whereas unmaterialised hazards detection varied by pathway, highest on pavements, moderate on roads, and lowest in cycle lanes, indicating strong contextual effects. These findings demonstrate that hazard perception develops with riding experience and is influenced by infrastructure, offering foundational evidence to inform targeted training interventions, infrastructure design adaptations, and regulatory measures to enhance e-scooter safety.