Exploration of Work Zone Crash Causes and Implications for Safety Performance Measurement Programs

Police crash reports are a fundamental data source for state work zone safety performance measurement programs.Past studies have often identified work zone crashes simply based on their location, and have not considered the causal role (if any) that the work zone played in the crash.This thesis identified crashes directly influenced by a work zone ("Directly Related" crashes) from a dataset of all crashes that were reported as occurring in a work zone ("Coded" crashes).This was done to assess the degree to which work zone activities appeared to actually influence the likelihood or severity of crashes within the work zone.Data was obtained from Virginia crash reports as well as the National Highway Traffic Safety Administration's (NHTSA) National Motor Vehicle Crash Causation Survey (NMVCCS) and Large Truck Crash Causation Study (LTCCS).Work zone crash reports were examined individually, and crash causal trends were identified within each dataset.The process of extracting Directly Related crashes and determining work zone crash causality from typical DOT databases (including Virginia's) is time-consuming and imprecise.In contrast, data obtained from the NMVCCS and LTCCS is more precise due to the studies' high level of detail.However, extracting necessary crash information from a NMVCCS or LTCCS crash report takes approximately 15 times longer than a Virginia crash report.The study found that there is a disparity between Coded crashes and Directly Related crashes.Only 23.0%, 53.8%, and 73.8% of the Virginia, NMVCCS, and LTCCS Coded crashes, respectively, could be directly related to the work zone.Implications of v the differences between Coded and Directly Related crashes on work zone safety performance measures are discussed.Several work zone crash causes in each of the four primary crash types (rear-end, angle, sideswipe -same direction, and fixed object -off road) were common across all datasets.Analysis identified several potential avenues where work zone safety performance measures could be influenced by the Virginia Department of Transportation (VDOT), including improved work zone traffic management planning, ingress/egress design, and traffic control placement.Datasets' analytical results were compared between the Virginia dataset and the NMVCCS and LTCCS datasets in order to determine if Virginia crash reports contained enough work zone related crash information in order to effectively and accurate develop crash trends and performance measures.This thesis found that Directly Related crash type proportions and crash cause proportions significantly varied between the Virginia dataset and the NMVCCS and LTCCS datasets.These differences are most likely the result of different levels of crash detail between the Virginia dataset and the NMVCCS/LTCCS datasets.Improvements to the Virginia crash forms are recommended in order to more accurately identify Directly Related crashes and their causes.