ESTIMATING BENEFITS FROM SPECIFIC HIGHWAY SAFETY IMPROVEMENTS

In the past thirty years, highway fatality rates have declined steadily because, most notably, of dramatic changes in motor vehicle design, passage of laws making seat belt use mandatory and driving while intoxicated a criminal offense, and educating the public through focused advertising campaigns. However, the practice of highway design has changed little. Standards, guidelines and warrants are based largely on the opinion of experts and the principles of geometric design. A key element in geometric design is providing the driver with an adequate sighting distance. For a given design speed, a highway is constructed with horizontal and vertical curvatures such that a driver has a sight line and sufficient time to recognize danger and to stop in a timely fashion. While the principle of adequate sight distance is embraced in practice, it has not been subject to in-depth scientific scrutiny. The objective of this research is to build upon the scientific framework for identifying hazardous highway locations already begun by these researchers by forming models for predicting the effects of traffic density and land use on highway safety. This will be accomplished by studying a small number of highway locations with varying background conditions. To focus the analysis and improve the quality of the results, study locations will be restricted to rural, two-lane highways. Effects attributable to these factors will be identified by comparing accident histories at sites where other background conditions are similar. At the present stage of development, models developed by these researchers prove helpful in identifying hazardous highway locations and in identifying contributing factors as to why the highways are deemed hazardous. With a predictive model, proposed measures to reduce both traffic demand and speed with changes in land use policy and highway design changes could be evaluated. The aim to develop these kind of predictive tools with our scientific framework of risk analysis is justifiable.