2019Traffic Injury PreventionRequires access

Modeling head-on crash severity with drivers under the influence of alcohol or drugs (DUI) and non-DUI

Pengfei Liu, Wei Fan

Open publisher page 11 citations

Abstract

Objective: The objective of this research is to identify and compare contributing factors to head-on crashes with drivers under and not under the influence of alcohol or drugs.Methods: The head-on crash data are collected from 2005 to 2013 in North Carolina from four aspects: vehicle, driver, roadway, and environmental characteristics. The final dataset includes 9,153 head-on crashes. A mixed logit model is developed to analyze the crash dataset involving drivers under and not under the influence of alcohol or drugs.Results: According to the obtained results, factors such as rural roadways, adverse weather, curve road, and high speed limit are among the most significant contributing factors to both head-on crashes with DUI and non-DUI. In addition, the results of this research demonstrate that high speed limit is found to be better modeled as random-parameters at specific injury severity levels for head-on crashes with DUI. Besides the factors mentioned above, dark light condition, old drivers, pickups, and motorcycles also significantly affect the severity of head-on crashes with non-DUI.Conclusions: The results of this study identify various factors that significantly affect the severity of head-on crashes with drivers under and not under the influence of alcohol or drugs. Also, the mixed logit model examines the heterogeneous effects and correlation in unobserved factors by allowing coefficients to be randomly distributed. The findings of this study call for more attention to head-on crashes and provide a reference for planners and engineers when developing and selecting countermeasures to reduce and/or mitigate head-on crashes.

About this research paper

What this paper is about

Objective: The objective of this research is to identify and compare contributing factors to head-on crashes with drivers under and not under the influence of alcohol or drugs.Methods: The head-on crash data are collected from 2005 to 2013 in North Carolina from four aspects: vehicle, driver, roadway, and environmental characteristics. The final dataset includes 9,153 head-on crashes. A mixed logit model is developed to analyze the crash dataset involving drivers under and not under the influence of alcohol or drugs.Results: According to the obtained results, factors such as rural roadways, adverse weather, curve road, and high speed limit are among the most significant contributing factors to both head-on crashes with DUI and non-DUI. In addition, the results of this research demonstrate that high speed limit is found to be better modeled as random-parameters at specific injury severity levels for head-on crashes with DUI. Besides the factors mentioned above, dark light condition, old drivers, pickups, and motorcycles also significantly affect the severity of head-on crashes with non-DUI.Conclusions: The results of this study identify various factors that significantly affect the severity of head-on crashes with drivers under and not under the influence of alcohol or drugs. Also, the mixed logit model examines the heterogeneous effects and correlation in unobserved factors by allowing coefficients to be randomly distributed. The findings of this study call for more attention to head-on crashes and provide a reference for planners and engineers when developing and selecting countermeasures to reduce and/or mitigate head-on crashes.

Why it matters

OpenAlex reports 11 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Objective: The objective of this research is to identify and compare contributing factors to head-on crashes with drivers under and not under the influence of alcohol or drugs.Methods: The head-on crash data are collected from 2005 to 2013 in North Carolina from four aspects: vehicle, driver, roadway, and environmental characteristics. The final dataset includes 9,153 head-on crashes. A mixed logit model is developed to analyze the crash dataset involving drivers under and not under the influence of alcohol or drugs.Results: According to the obtained results, factors such as rural roadways, adverse weather, curve road, and high speed limit are among the most significant contributing factors to both head-on crashes with DUI and non-DUI. In addition, the results of this research demonstrate that high speed limit is found to be better modeled as random-parameters at specific injury severity levels for head-on crashes with DUI. Besides the factors mentioned above, dark light condition, old drivers, pickups, and motorcycles also significantly affect the severity of head-on crashes with non-DUI.Conclusions: The results of this study identify various factors that significantly affect the severity of head-on crashes with drivers under and not under the influence of alcohol or drugs. Also, the mixed logit model examines the heterogeneous effects and correlation in unobserved factors by allowing coefficients to be randomly distributed. The findings of this study call for more attention to head-on crashes and provide a reference for planners and engineers when developing and selecting countermeasures to reduce and/or mitigate head-on crashes.

Key concepts: Crash, Driving under the influence, Speed limit, Affect (linguistics), Transport engineering, Logistic regression, Poison control, Environmental health

Related papers

Back to paper searchBrowse research topicsOriginal source
Modeling head-on crash severity with drivers under the influence of alcohol or drugs (DUI) and non-DUI — Research Paper | ScholarLens