Alternative Ordered Response Frameworks for Examining Pedestrian Injury Severity in New York City

This paper focuses on identifying the appropriate ordered response structure that is better suited to modeling pedestrian injury severity. The alternative ordered response approaches considered for the empirical analysis include: ordered logit model (OL), generalized ordered logit model (GOL) and latent segmentation based ordered logit model (LSOL). The GOL model relaxes the restrictive assumption by allowing for exogenous variable impacts on the threshold parameters in the standard ordered logit structure. Again, the LSOL model allows for differential impact on the alternatives by segmenting the pedestrian crash population into various segments with segment specific ordered logit parameters. The performance and strength of the formulated injury severity models are tested based on the “New York City (NYC) Pedestrian Research Data Base” for the year of 2002 through 2006. To our knowledge, the study provides a first of its kind exercise to identify the preferred ordered model for examining pedestrian injury severity. The model estimation results clearly highlights the presence of segmentation based on the location of pedestrian accidents; locational attributes that affect the allocation of pedestrians into these segments include: regional county, functional classification of roadway, pedestrian location on roadway, number of travel lanes and number of parking lanes in the roadway system. The key factors influencing pedestrian injury severity are weather condition, lighting condition, vehicle types, pedestrian age and season. Overall, the results of the empirical analysis provide credence to the belief that LSOL model is a preferred ordered framework choice to accommodate population heterogeneity in the context of pedestrian injury severity.

  • Supplemental Notes:
    • This paper was sponsored by TRB committee ANF10 Pedestrians.
  • Corporate Authors:

    Transportation Research Board

    500 Fifth Street, NW
    Washington, DC  United States  20001
  • Authors:
    • Yasmin, Shamsunnahar
    • Eluru, Naveen
    • Ukkusuri, Satish V
  • Conference:
  • Date: 2013

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: References; Tables;
  • Pagination: 22p
  • Monograph Title: TRB 92nd Annual Meeting Compendium of Papers

Subject/Index Terms

Filing Info

  • Accession Number: 01477594
  • Record Type: Publication
  • Report/Paper Numbers: 13-3987
  • Files: PRP, TRIS, TRB, ATRI
  • Created Date: Apr 4 2013 4:53PM