Do motorists select safe routes over efficient routes? An empirical study using travel trajectory data

Studies have highlighted the importance of traffic safety concerns in travelers' mode choices. However, little attention has been given to route choices between safety and efficiency. This study aims to investigate motorists' safety route choice behavior using vehicle trajectory data collected by the Automatic License Plate Recognition (ALPR) system in Changsha, China. A paired scenario comparison (PSC) method is proposed to examine motorists' safety route choice behaviors by contrasting the choices under two scenarios across a large-scale city area. Specifically, four scenarios are considered, including daytime/nighttime, work/non-work, livelihood/non-livelihood, and short-distance/long-distance travel. Motorists' route choice preferences are hypothesized to differ between paired scenarios. Motorists' route choices are depicted by the Group Random Parameters Logit (GRPL) model with a set of route safety and efficiency attributes. Hypotheses are tested through an overall model validation and a series of scenario-specific model validations. Results demonstrate the significance of motorists' priority of the routes with safety attributes (e.g., fewer intersections, wider roads) during nighttime trips or for non-livelihood purposes. However, evidence for work/non-work and trip distance effects remains inconclusive. This study provides empirical support for safety-oriented route planning and navigation systems.