On the significance of omitted variables in intersection crash modeling

Advances in safety research – trying to improve the collective understanding of motor vehicle crash causes and contributing factors – rest upon the pursuit of numerous lines of research inquiry. The research community has focused considerable attention on analytical methods development (negative binomial models, simultaneous equations, etc.), on better experimental designs (before–after studies, comparison sites, etc.), on improving exposure measures, and on model specification improvements (additive terms, non-linear relations, etc.).