Development of usage guidelines for speed display units in school zones
University of New Brunswick
Road authorities are under increasing pressure from parents and the public to install speed display signs in school zones as a perceived means to improve safety for students. While isolated studies have shown that many of these supplemental countermeasures yield quantifiable changes in driver behaviour, little empirical evidence exists regarding collision experience within designated school zones to quantify whether a problem actually exists. This research analyzed collision history data for delineated school zones/areas from a sample of urban and rural locations in New Brunswick with the goal of developing usage guidelines for speed displays that can assist municipalities to make an objective decision regarding the implementation of this countermeasure. Results from the collision analysis showed that only 21% of urban and 29% of rural schools are performing statistically worse than the group as a whole. Comparison of yearly collisions found only 2% of schools experienced statistically more collisions than a comparable road without a school. Results indicate that schools in both urban and rural locations are performing better than expected. A collision prediction model was developed using a Zero-Inflated Negative Binomial regression that relates collision data to explanatory school zone/area characteristics. Variables found to have the most significant effect on collisions within school zones/areas include road volume, presence of a signalized intersection, number of unsignalized intersections present, and number of through lanes. The resulting model subsequently provides a warrant system designed to evaluate whether a specific school site would justify the installation of a speed display unit that is based on contributing variables to school zone collisions.