Subjective and objective risks consideration in modeling highway safety

Alberto Miguel Figueroa-Medina, Purdue University


The study of highway crash occurrence and its causation has been a prominent research subject in recent decades due to its significant impact and cost to society. Highway crashes are random events that are caused by many and very different contributing factors. The principal factors are categorized as roadway, vehicle, and driver factors and their relationships with crashes are extremely complex. Extensive research has resulted in prediction models that separately link the roadway characteristics with the speed or the crash frequency (or rate). It is accepted that a link exists between the objective risk and the subjective risk that is determinant in the likelihood of a crash, but their relationship has not been fully researched. This dissertation presents the results of a study that measured and evaluated the risk perception of drivers using roadway characteristics, free-flow speeds, and crash data collected on four-lane highway sections in Indiana. The results of the risk perception study shed light on the relationship between the objective risk and the subjective risk and provided an understanding of the individual effects of the roadway and driver characteristics on the risk perception. The goal of this dissertation is to present the results of an exploratory analysis of the relationship between roadway characteristics, driver behavior as influenced by the perceived risk, and safety. An advanced econometric model was developed using a system of simultaneous equations that links the speed, the crash rate, and the roadway characteristics. The model provides a valuable contribution to the field of highway safety evaluation. The model results confirm that increasing speed is related to increasing crash rates; and that drivers, on average, can perceive the risk of crash on the road based on the roadway characteristics, and consequently adjust their behavior by modifying their speed. The use of a simultaneous equations approach to calibrate the proposed system of relationships, aside from being the correct formulation to model the endogenous relationship between speed and crashes, proved to be consistent with rational driver behavior and improved the identification of speed and crash rate factors when compared with the single-equation modeling approaches.




Tarko, Purdue University.

Subject Area

Civil engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server