RESEARCH ARTICLE
Examining the Severity of Traffic Barriers Crashes, Mixed Model with Observed Heterogeneity
Mahdi Rezapour1, *, Khaled Ksaibati1
Article Information
Identifiers and Pagination:
Year: 2022Volume: 16
E-location ID: e187444782208200
Publisher ID: e187444782208200
DOI: 10.2174/18744478-v16-e2208200
Article History:
Received Date: 10/2/2022Revision Received Date: 19/4/2022
Acceptance Date: 26/5/2022
Electronic publication date: 21/10/2022
Collection year: 2022
open-access license: This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International Public License (CC-BY 4.0), a copy of which is available at: https://creativecommons.org/licenses/by/4.0/legalcode. This license permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Abstract
Introduction:
Due to the high involvement of traffic barriers in the severity of crashes, extensive efforts have been made to find factors to those crashes. In this study, the mixed logit model has been recognized and employed for modeling traffic barriers crash severity.
Methods:
The method has shown an improvement over the standard logit model, which assumed the impacts of predictors are fixed across crash observations. However, most past studies assume constant distributional means across various crash observations despite the efforts. In this study, the random parameter model was extended to incorporate the heterogeneity in the taste of random parameters based on other observed factors. The consideration addresses the limitation of the standard mixed model, constraining the random effect means to be constant across all observations. In this study, the heterogeneity in taste highlights a significant difference across subpopulations of barrier crash severity based on various factors. The results of the goodness of fit also highlight significant improvements in model fits, moving from standard logit to the mixed and the mixed models with heterogeneity in tastes.
Results and Discussion:
The results highlight that, for instance, the means of the random parameters of gender varies across crash population based on shoulder width, and average annual daily traffic (AADT), while the impact of the mean of the random parameter of AADT varies based on truck traffic.
Conclusion:
Driver's restrain condition, rollover type of crashes, posted speed limit, and citation record were some of the factors that their effects on the severity of crashes were found to be fixed.
