CALAX: A collision taxonomy for research and traffic records

This article describes a collision taxonomy that has three main advantages. First, its coding reliability was established; coders independently coding the same cases achieved 96% agreement. Second, it has a hierarchical structure, enabling the user to choose one of three levels of detail; there are 46 collision types at the most detailed level. Third, it captures the role of the individual vehicle, a valuable feature for suggesting the driver problems involved in an accident. The taxonomy was developed after reviewing other systems and incorporates some of their better features. It uses schematic diagrams to facilitate ease of learning and use, and it was developed for use by researchers and for traffic record systems. The value of the taxonomy for suggesting driver problems in accidents was demonstrated with data indicating alcohol impairment and age effects.