Author ORCID Identifier
Year of Publication
Doctor of Philosophy (PhD)
Dr. Reginald Souleyrette
Highway safety management aims to prevent crashes and reduce the resulting frequency and severity within the limit of available resources. The identification of potentially hazardous sites and investment on safety treatments have been fundamental to fulfill this goal. However, having many highway safety improvement projects in hand, safety professionals need to evaluate several different alternatives and allocate the limited funds to the ones that would provide the highest return on investment. Hence, prioritizing the safety projects based on their potential to achieve the greatest safety benefit is crucial. Ineffective prioritization can distribute funds to locations with less potential for improvement while higher-risk sites remain untreated.
Prior to the widespread use of the Highway Safety Manual (HSM), highway safety analysts prioritized candidate safety improvement projects using simplistic safety metrics based only on past performance (e.g., crash history, rates, and costs). However, these metrics lack precision and are limited by several methodological weaknesses. Published in 2010, the HSM provides a comprehensive guideline for evaluating safety improvements that facilitates the use of advanced safety performance measures including “Excess Expected Crashes (EEC)”. This metric is dependent on two estimates: expected crashes by Empirical Bayes (EB) method and predicted crashes by Safety Performance Functions (SPFs). It is obtained by taking the difference between these estimates and better reflect a site’s safety improvement potential. Nowadays, several agencies and state departments of transportation use EEC for prioritizing projects.
While EEC is a preferred measure, it comes with a few limitations. The use of EEC alone may not identify all sites with promise, as it purposefully ignores the real maximum potential reduction, that is, to zero. Further, if SPFs (and EEC) are based on combined crashes of all severities, the focus may be placed on the less serious crashes that comprises the substantial proportion. However, focusing only on the most severe crashes often leads to small sample size problems that hinders the development of meaningful prediction models. Moreover, EEC compares the safety performance of a site to the average location for that roadway type and traffic volume but fails to represent the magnitude of the overall number of crashes occurring at that site, more preciously EB estimates. Additionally, the ever-present need for choosing the most appropriate SPFs that can balance between the model quality and available data remains.
The research conducted in this dissertation aims to propose a more comprehensive approach to prioritize the potential for safety improvement of proposed projects. To illustrate the concepts developed in this work, a case study (a project prioritization scheme named Strategic Highway Investment Formula for Tomorrow (SHIFT) by Kentucky Transportation Cabinet (KYTC)) is presented using crash and roadway data from state-maintained roads in Kentucky. This scheme used EEC of total crashes for project ranking. Their desire to further improve the safety ranking methodology manifested the impetus for this research.
This dissertation offers recommendations regarding the choice of SPFs and integration of crash severity and EB estimate for project prioritization. It also offers a novel technique of considering severity by incorporating future goals of fatalities into the project prioritization metric-EEC. Finally, it presents a framework for developing a multifactor safety scoring technique using Kentucky data and recommends replacing the dependency on EEC only with this new safety score for prioritizing safety projects. In the process of conducting the research, a tool was developed that automates the safety score estimation and ranking with efficiency. This technique and tool can be customized according to a jurisdiction’s safety needs and therefore, can be used for any state or country’s safety investment.
Digital Object Identifier (DOI)
Tanzen, Riana, "OPTIMIZING THE POTENTIAL OF HIGHWAY SAFETY INVESTMENT" (2022). Theses and Dissertations--Civil Engineering. 126.