Identifying street design elements associated with vehicle-to-pedestrian collision reduction at intersections in New York City

Abstract

Objective: We evaluated associations between the installation of eleven street design elements, between 2007 and 2015, and subsequent changes in vehicle-to-pedestrian collisions in New York City.

Methods: Collision data were from Accident Location Information System in the New York State Department of Transportation. Safety improvement projects at 118 intersections were reviewed and their implemented street design elements were identified. First, we assessed potential regression-to-the-mean effects using historic trends of pedestrian collision count at the intersection project locations. Second, we used a two-group pretest-posttest design to assess individual element's associations with pedestrian collision reduction after installations. Pedestrian collision count and pedestrian- and vehicle-based pedestrian collision rates were examined. Third, regression trees were used to classify the intersections with design elements as independent variables for the target variables of collision outcomes, to identify street design element combinations associated with pedestrian collision reductions.

Results: Treatments with pedestrian refuge island or pedestrian plaza had reductions in pedestrian collision count and pedestrian-based collision rate while their comparisons had no changes. Treatments with pedestrian refuge island had a larger reduction in pedestrian collision when combined with lane removal or narrowing. Treatment with curb extension or pedestrian plaza had reductions in vehicle-based pedestrian collision rate while their comparisons had no changes. Other studied elements showed no, small, or insignificant associations with post-project pedestrian collision reductions.

Keywords: Countermeasure; Pedestrian accident; Pedestrian crash; Quasi-experimental design; Regression tree; Traffic calming.