Potential Uses of Reduced Datasets from the Roadway Information Database: A White Paper

While the second Strategic Highway Research Program (SHRP2) Roadway Information Database (RID) was designed to link to SHRP2 Naturalistic Driving Study data, the level of coverage and accuracy of its mobile data, combined with the inclusion of supplemental data from existing sources, makes the RID a powerful stand-alone database as well. The RID as a stand-alone database is particularly useful for research efforts that seek to relate the safety performance of road segments and/or intersections, measured by the frequency, type, and severity of crashes, to the segment and intersection traffic, traffic control, and geometric characteristics. This white paper describes potential highway safety research questions that can be addressed using five new RID reduced datasets created by the FHWA Office of Safety Research and Development. These five new datasets are (1) average annual daily traffic, (2) intersection widths, (3) intersection crashes, (4) curve crashes, and (5) homogenous segments. Chapter 1 of the white paper provides an overview of the RID, a discussion of the value of the RID as a stand-alone database, and a definition for a RID reduced dataset. Chapter 2 provides overviews of the five new reduced datasets along with previously developed reduced datasets that are complementary to the new ones. Chapter 3 contains potential highway safety research questions that can be addressed using the five new datasets independently or in combination with previously developed reduced datasets. Potential research topics span seven topic areas: (1) crash predictions, (2) safety performance impacts of horizontal curve features, (3) safety performance impacts of intersection features, (4) safety performance impacts of access management, (5) risk factors for systemic safety analysis, (6) crash assignments, and (7) driver awareness of signalized intersections when entering urban areas (following rural travel). Chapter 4 provides a summary and conclusions.