This article describes a study in which the effectiveness of traffic control devices used in urban work zones was evaluated using data for sites located in Virginia. Regression models were developed relating work-zone accident rates with prior accident rates and types of control devices used. These regression models were developed for both two-lane and multi-lane highways. The traffic control devices evaluated were those used at the sites selected but include most of the devices commonly used at urban work zones (barricades, cones, flashing arrows, static signs, flagmen and temporary pavement markings). Use of advance warning signs was not considered in this study. Results indicate that the accident rate during periods of maintenance and rehabilitation work for multi-lane highways is highly dependent on the accident rate just prior to the work period. (A)
Abstract