Development of a Crash Risk-Scoring Tool for Pedestrian and Bicycle Projects in Oregon

Abstract

Methods for identifying and prioritizing high-crash locations for safety improvements are generally crash-based. There are fewer reported crashes involving non-motorized users and, in most states, reported crashes must involve a motor vehicle. This means that minor, non-injury events are not reported and those crashes that are reported tend to be more severe. Selecting projects based only on crash performance is sometimes limiting for these crash types and predicting where these crashes will occur next is also a challenging task. An alternative to crash-based selection is to develop risk-based criteria and methods. This paper presents the results of a research effort to develop a risk-scoring method with weights derived from data for use in project screening and selection in Oregon. To develop the risk model, data were collected from 188 segments and 184 intersections randomly selected on both state and non-state roadways. Geometric, land use, volume, and crash data were collected from Google Earth, EPA’s Smart Location Database, and the Oregon Department of Transportation crash database from 2009 to 2013. The sample included 213 bicycle and pedestrian crashes on the segments and 238 at intersections. Logistic regression models were developed and the outputs used to create pedestrian and bicycle risk-scoring tools for segments and intersections. The risk-scoring tool was applied to safety projects identified in the 2015 All Roads Transportation Safety (ARTS) project lists from Oregon. The risk scores for the case study applications aligned reasonably well with the project’s benefits-costs estimates.