Analysis of Data Selection and Inclusion for Highway Safety Manual Model Calibration

Abstract

The application of the American Association of State Highway and Transportation Officials’ (AASHTO) Highway Safety Manual (HSM) to Louisiana roads is a key component to the Louisiana Department of Transportation and Development’s (LA DOTD) plan to improve safety on state highways and reach the goal of “Destination Zero Deaths.” To apply the HSM the research sought to develop state-specific HSM calibration factors for eight facility types. During the development process, the data-intensive computational procedure followed to compute the calibration factors revealed numerous issues associated with the inclusion and definitions of various data elements required by the HSM. These included coding errors, missing required data elements in the Louisiana roadway and crash databases, and varying definitions of which specific crashes should or should not be included in the sample. Because of this, some of the resulting factors were unexpected, in particular, those for urban three-lane and urban five-lane highways which were lower than anticipated. To investigate the effect of including or excluding various data elements and varying crash definition assumptions, the HSM calibration factors were developed using a series of computational iterations in which the amount of data and assumed crash conditions were varied from one iteration to the next. The overall results of this thesis demonstrate the extent of the variability and sensitivity of HSM calibration factors to the inclusion of data that may or may not be included in roadway databases and how crashes that occur within various distances away from intersections are included or excluded.