An Empirical Assessment and Investigation of the Driver Injury Severities in Rain-Related Rural Single-Vehicle Crashes Using Mixed and Latent-Class Logit Models

Due to slippery road surface together with the limited visibility, single-vehicle crash during rain, especially the one occurred in the rural area, is more likely to result in driver incapacitating injury or even fatality. A two-year crash dataset including all rain-related rural single-vehicle crashes in four South Central states, i.e., Texas, Arkansas, Oklahoma, and Louisiana, from 2011 to 2012 were selected in this paper to analyze the impacts of the risk factors on driver injury severity. Mixed multinomial logit model and latent class multinomial logit model were both developed using the same dataset. Several parsimony indices including Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), as well as McFadden pseudo r-squared, are calculated for each model to evaluate their performances. Results showed that choosing normal distribution as the prior for random parameters could best increase goodness-of-fit of the mixed logit model. In addition, the two-class latent class model also showed superiority when compared to the three- and four-class models. Finally, a careful comparison between these two models was conducted, and the results indicated that the latent class logit model behaves better in analyzing the aforementioned dataset in this study. Model estimation results showed that curve, on grade, signal control, multiple lanes, pickup, straight, drug/alcohol impaired, and seat belt not used have adverse impacts on driver injury severity in the two models. On the other hand, wet, male, semi, and young have favorable effects on injury outcomes. This study provides insightful understandings of the effects of these attributes on rain-related single-vehicle crashes and beneficial references for developing effective countermeasures for severe crash prevention.

• Supplemental Notes:
  • This paper was sponsored by TRB committee ANB20 Standing Committee on Safety Data, Analysis and Evaluation.
• Authors:
  • Li, Zhenning
  • 0000-0002-0877-6829
  • Ci, Yusheng
  • Chen, Cong
  • Zhang, Guohui
  • Wu, Qiong
  • Qian, Zhen (Sean)
  • Prevedouros, Panos D
  • 0000-0001-6983-2321
  • Ma, Tienwai
  • 0000-0002-1826-8844
• Conference:
  • Transportation Research Board 97th Annual Meeting
  • Location: Washington DC, United States
  • Date: 2018-1-7 to 2018-1-11
• Date: 2018

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; References;
• Pagination: 7p

Subject/Index Terms

• TRT Terms: Crash data; Injury severity; Logits; Rain; Rural areas; Single vehicle crashes
• Geographic Terms: Arkansas; Louisiana; Oklahoma; Texas
• Subject Areas: Highways; Safety and Human Factors;

Filing Info

• Accession Number: 01663078
• Record Type: Publication
• Report/Paper Numbers: 18-06244
• Files: TRIS, TRB, ATRI
• Created Date: Mar 20 2018 5:08PM