Assessing Critical Factors Associated with Bicycle Collisions at Urban Signalized Intersections

Since it is an important process to understand what factors are related with bicycle crashes at urban signalized intersections for improving the safety of bicyclists and guiding safe designs of urban signalized intersections, this study developed prediction models for bicycle crash occurrences at signalized intersections using numerous potential variables relating with bicycle crashes by conducting field survey at 151 intersections in Inchon, Korea. After careful application and assessment of statistical model (of the ones fitted) for bicycle-related crashes, this study found Poisson regression model could be suitable for estimating the probability of bicycle crashes at intersection because the coefficient of over-dispersion was close to zero. Based on the analysis of the parameters estimated in models including the primary and alternative models, significant explanatory factors (and their direction of association) were selected as follows: average daily traffic volume (+), presence of bus stops (-), sidewalk widths (-), number of driveways (+), presence of speed restrict devices (-), presence of crosswalks (+), and industrial land use (+). With respect to suggestions for future bicycle safety research, there is a need to include additional driver, geometric, and operational features in an analysis. Educational approaches or improvement of roadway designs should be also performed to encourage people to safely use bicycles as an alternative travel mode. Finally, this study believe that the levels of safety of bicycle travels at currently existing or future intersections can be estimated through bicycle crash prediction models and efficient countermeasures can be implemented to decrease crash rates and reduce socio-economic loss.

Language

  • English

Media Info

  • Media Type: DVD
  • Features: Figures; References; Tables;
  • Pagination: 17p
  • Monograph Title: TRB 87th Annual Meeting Compendium of Papers DVD

Subject/Index Terms

Filing Info

  • Accession Number: 01103548
  • Record Type: Publication
  • Report/Paper Numbers: 08-0832
  • Files: TRIS, TRB
  • Created Date: Jun 24 2008 11:42AM