Development of systemwide pedestrian safety performance function using stratified random sampling and a proxy measure of pedestrian exposure

The lack of pedestrian counts at a systemwide level prompts the need to find other innovative ways of assessing pedestrian traffic crash risks using proxy measures of exposure. This study aims to formulate the methodology for developing pedestrian safety performance functions (SPF) using the proxy measure of pedestrian exposure and stratified random sampling. The case study was all urban intersections in Michigan State that comprise of collector and arterial roads. The stratified random sampling strategy was deployed to select the sample which is representative of all urban intersections in the state of Michigan. Factor analysis was used to develop a proxy measure of pedestrian exposure at urban intersections using a walkability measure (walk score), among other factors. The performances of various count models were compared using the goodness of fit measures based on the Akaike’s Information Criterion (AIC), Bayesian Information Criterion (BIC), and Vuong test. The final pedestrian SPFs was formulated using the Zero-Inflated Poisson (ZIP) model with average annual daily traffic (AADT) at a major approach, AADT at the minor approach, and a proxy measure of pedestrian exposure. The proposed methodology in this study can benefit transportation agencies that have embarked on systemwide planning of pedestrian facilities to improve the safety of pedestrians but lack systemwide analytical tools and pedestrian counts to make data-driven decisions.

  • Record URL:
  • Availability:
  • Supplemental Notes:
    • © 2020 Informa UK Limited, trading as Taylor & Francis Group. Abstract reprinted with permission of Taylor & Francis.
  • Authors:
    • Kwayu, Keneth Morgan
    • Kwigizile, Valerian
    • Oh, Jun-Seok
  • Publication Date: 2020-10

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01765790
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Dec 31 2020 3:01PM