Safety of high speed expressway signals: a comparison of classical and empirical Bayes methods
High speed expressways are becoming increasingly common as two-lane roads are improved to handle suburban traffic growth. Characterized by at-grade intersections and at least two lanes of traffic in each direction, these facilities are separated by a median and commonly have speed limits of 50 mph or greater. As traffic levels increase, stop controlled intersections are typically signalized to reduce delay or enhance safety. It is the safety performance of these signals that is the focus of this thesis. This thesis reports on a study investigating the safety benefit of signalizing intersections of high speed divided expressways. Cross classification, matched (yoked) pairs, before and after, and empirical Bayes (EB) analyses were conducted on 50&55-mph and 55-mph only intersections comparing unsignalized and signalized intersections. The results show that, generally, signalized intersections have a higher crash rate and lower costs per crash. However, in the before and after analysis (intersections that were signalized between 1994 and 2001), the after period experienced lower crash rates with higher costs per crash than before signalization. In the EB analysis, the crash rates changed from the before and after analysis (from 11.7% decrease to an 18.9% increase at the 50&55-mph intersections and from 31.4% decrease to a 6.8% decrease at the 55-mph only intersections).