Using Structural Equation Modeling to Measure Single-Vehicle Crash Severity

• Activity Number: 437
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Government Statistics Section
• Author(s): Xiao Qin*+
• Keywords: Highway Safety ; Injury Severity ; Structural Equation Model

Abstract

Injury severity and vehicle damage are two important indicators of the level of crash severity. Factors such as driver and highway characteristics, environmental factors, vehicle types and roadside objects may be directly or indirectly related to the consequence of a crash, i.e. injury severity and vehicle damage. All these factors interact in complicate ways so that the relationship among them are difficult to be identified.

The structural equation modeling (SEM) is adopted to unravel the complex relationships among variables because it can handle endogenous and exogenous variables simultaneously; and furthermore it can include latent variables to bridge the gaps between dependent and explanatory variables. In this study, to estimate the relationships between exogenous factors and crash severity, three structural equation models with one, two and three latent variables were developed and compared. Based on the goodness-of-fit and model predictive power, the structural equation modeling with two latent variables outperformed the other two. Additional insights about the model selection were also provided through the development and comparison of the three models.