A Spatiotemporal Random Parameters Model to Formulate Impaired-Driving Crash Likelihood

Unobserved heterogeneity across space, time, and crash type in crash frequency modeling has been recognized as a critical problem in crash frequency modeling. In this study, five hierarchical Bayesian models with different spatiotemporal interactions are separately developed to address this issue. The yearly county-level alcohol/drug impaired-driving related crash counts data of three different injury severities including minor injury, major injury, and fatal injury in Idaho from 2010 to 2015 is selected for analysis. Based on posterior predictive check and DIC results, the main effect model with the interaction of structured temporal effects and unstructured spatial effects has the best model performance and is selected for further analysis. Total population, unemployment rate, and the percent of 25 years and older with a bachelor's degree or higher are significantly associated with the frequencies of any of the three crash types. In addition, the proportion of male is found to significantly influence the crash counts of fatal crashes. However, pavement condition and number of lanes are not found significantly associated with crash frequencies for all the three crash types and are removed from the final model. Results also suggest that ignoring temporal and spatial heterogeneity may result in biased parameter estimates and incorrect inferences.

  • Supplemental Notes:
    • This paper was sponsored by TRB committee ANB50 Standing Committee on Alcohol, Other Drugs, and Transportation. Alternate title: A Spatiotemporal Bayesian Hierarchical Approach to Analyze Alcohol/Drug Impaired-Driving Related Traffic Crashes by Severity.
  • Corporate Authors:

    Transportation Research Board

  • Authors:
  • Conference:
  • Date: 2019


  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures;
  • Pagination: 5p

Subject/Index Terms

Filing Info

  • Accession Number: 01698285
  • Record Type: Publication
  • Report/Paper Numbers: 19-05183
  • Files: TRIS, TRB, ATRI
  • Created Date: Dec 7 2018 9:50AM