Evaluating Contributing Factors to Collision Types Through Discrete Choice Analysis
Author: Dejan Dudich
Publisher:
Published: 2016
Total Pages: 94
ISBN-13:
DOWNLOAD EBOOKWhile there have been several efforts to understand large-truck crashes, the relationship between crash factors, crash severity and collision type is not clearly understood. Past studies have utilized different statistical or econometric models to predict the manner of collision at intersections, yet not much attention has been paid to the factors that lead to injury severity by different types of collisions on state and interstate highways. Studying collision types is crucial when identifying potential safety improvements for state and interstate systems. In this study six collision types are explored they are: angled collisions, fixed object collisions, rear end collision both vehicles moving forward, rear end collisions on moving vehicle, sideswipe collision same direction and sideswipe collisions different directions. With these in mind, the aim of this research is to perform exploratory analyses of large truck-involved crashes through the use of advanced econometric techniques that can shed insights on the factors influencing crashes by collision type. Namely, this research utilizes the mixed multinomial logit model to uncover the effects of unobservable factors (unobserved heterogeneity) across crash observations underlying the data generating process. The results of this thesis indicate that complex interactions of various human, vehicle, and road-environment factors due in fact contribute and that some of the model variables varied across observations, validating the choice of the mixed multinomial logit model and separation of data by collision type.