| Rural highways represent the largest single class of highways in the United States, and they account for approximately 80 percent of all paved highways. Two-lane highways account for more than 85% of all rural highways, and fatal crashes nationwide on these roadways totaled 19,055 in 2000. Transportation safety managers are increasingly responsible for identifying and investing public monies in highway projects that result in the greatest reduction of fatalities, injuries and property damage resulting from motor vehicle crashes.; Prior to the implementation of any given highway safety countermeasure, safety managers need a repeatable and objective process that evaluates the expected effect on highway safety from these highway projects. Objective methods for countermeasure selection are limited. Safety managers could use the results from literature reviews or locally funded research in selecting projects, but these approaches face technical deficiencies such as constraints due to right-of-way or environment features, are costly, and do not account for local or regional influences. In addition, many studies only provide insight of the implementation and effects of a single safety countermeasure at a time, whereas in reality multiple countermeasures could be applied.; The approach presented in this thesis proposes an analytical framework to assist highway safety managers in selecting highway safety countermeasures with the greatest safety benefit. This methodology, which relies on Bayesian statistical methods, meta-analytical methods, and subjective engineering evaluations, is termed the Bayesian Safety Assessment Framework (B-SAF) and is used to identify, assess, and rank safety countermeasure effectiveness in regional highway safety programs. The B-SAF methodology is sufficiently general in that it can be applied for managing safety in local, regional, or state jurisdictions. Additionally, this B-SAF process is beneficial, as safety countermeasures that have yet to be applied in their region of interest can be evaluated without being physically implemented beforehand. |
