A Modified Rank Ordered Logit Model to Analyze Injury Severity of Occupants in Multi-Vehicle Crashes

Date of Award

Summer 2015

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Civil & Environmental Engineering

Program/Concentration

Civil Engineering

Committee Director

Rajesh Paleti

Committee Member

Mecit Cetin

Committee Member

Filmon Habtemichael

Call Number for Print

Special Collections LD4331.E542 B64 2015

Abstract

Research on methodological approaches to model injury severity has become important in order to prevent crash fatalities. Roadways have always been dangerous, with significant congestion and uncontrollable variables, such as weather, coupled with road users' decisions including seat belt and alcohoVdrug use. These decisions can influence the severity of injuries sustained by vehicle occupants in the event of a crash. Few studies have examined the potential impacts of comprehensive roadway user injury severity profiles of all occupants ( drivers as well as passengers) in all vehicles involved in a crash. The focus of past studies has been on examining the factors that influence the severity outcomes of the driver or the most severely injured occupant in a crash. While these studies have been central to several policy decisions that have enhanced road safety, they are not capable of modeling the severity outcomes of multi-vehicle crashes and can thus lead to wrong policy implications.

The current study addresses this limitation head-on by developing a simultaneous model of severity outcomes for all people involved in a crash. Specifically, a Modified Rank Ordered Logit (MROL) methodology that can determine the relative order of occupant injury severity as well as the actual injury severity was developed. The final model quantifies the effects of several key occupant, vehicle, and accident level variables on four possible levels of injury severity. The presence of crash-level unobserved factors that can moderate the influence of these key variables was also accounted for, by allowing random parameter heterogeneity in the parameter estimates of the final model. The results indicate the presence of strong crash level unobserved factors that influence all severity outcomes as well as random heterogeneity in the effect of key covariates including occupants' gender, speed limit, and seating position. The performance of the MROL model was compared with the traditional mixed multinomial logit (MMNL) model that is the most commonly used model for safety analysis in both research and practice.

Overall, the results demonstrate the superior predictive ability of the MROL model in comparison to the MMNL model. The traditional MMNL model performed satisfactorily in terms of replicating the simple shares of different injury severity levels across all occupants. However, the performance of the MMNL model dropped significantly when the observed and predicted shares were compared for combinations of injury severity levels among crashes involving multiple occupants. Lastly, the policy implications of the MROL and MMNL models were compared using elasticity calculations. The results from elasticity analysis suggest that, not only does the MROL model provide superior data fit, it also provides considerably different elasticity effects, particularly for occupant level factors in comparison to the MMNL model.

Rights

In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).

DOI

10.25777/gc10-q831

This document is currently not available here.

Share

COinS