Evaluation of Surrogate Measures for Pedestrian Safety in Various Road and Roadside Environments

This report presents an investigation of pedestrian conflicts and crash count models to learn which exposure measures and roadway or roadside characteristics significantly influence pedestrian safety at road crossings. Negative binomial models were estimated for pedestrian conflicts and crash counts except for fatal and incapacitating crashes for which binary logistic models were estimated. Also models for predicting highest severity at a location were estimated using an ordered proportional odds (PO) technique. Pedestrian counts and conflicts data were collected using a variation of the Swedish Traffic Conflict Technique (TCT) at 100 locations throughout Connecticut. Pedestrian crash data for the latest available three years (2009, 2008, and 2007) were collected from the Connecticut Crash Data Repository (CTCDR). The results show that minor and serious conflicts are marginally significant in predicting total pedestrian crashes together with crossing distance and building setback. This suggests that these conflicts, when observed over a longer period of time, may be a good surrogate for crashes in analyzing pedestrian safety. Greater crossing distance and small building setbacks are both found to be associated with larger numbers of pedestrian-vehicle crashes. This latter effect is not expected, since vehicle speeds are expected to be lower in areas where the building setback is small. This factor may account for the greater pedestrian activity and more complex interactions in such areas. Further research aimed at identifying a minimum length of time for accurate estimation of pedestrian volume and conflicts to relate to crashes is the subject of continuing investigation by the authors.