Comprehensive Evaluation of Wisconsin Roundabouts Volume 2: Traffic Safety

While roundabouts are still fairly new in the U.S. and Wisconsin, their safety benefits have been studied with varied results. For this study, researchers analyzed 24 Wisconsin roundabouts that were built in 2007 or before. Three years of before and after crash data were gathered as well as geometric and volume data. An Empirical Bayes (EB) analysis was used to examine the safety benefits for total crashes and injury (K, A, B, C) crashes. A simple before-and-after crash analysis was also completed to analyze specific types of injury crashes for each roundabout. The EB analysis was performed using Safety Performance Functions (SPFs) from both the Highway Safety Manual (HSM) and Wisconsin specific data. The results using HSM and Wisconsin values were very similar, supporting the consistency between SPFs. Considering only the HSM SPFs, researchers found mixed results for total crash frequency but a significant decrease in crash severity. Wisconsin roundabouts had an unbiased estimate of a 9 percent decrease in total crashes with a standard error of 6 percent. National numbers similarly show decreases in total crashes. Wisconsin roundabouts showed a significant 52 percent decrease in injury crashes. Roundabouts nationwide are also experiencing a significant decrease in severe crashes. When evaluating predictor variables, the speed limit of the approaches did not show a significant impact on the safety of the roundabout. While multi-lane roundabouts seem to be safer than single lane roundabouts when considering combined injury crashes, single lane roundabouts saw the largest decrease in total crashes. Two-way stop controlled (TWSC) intersection conversion to a roundabout had the highest safety benefit as compared to all-way stop controlled (AWSC) and signalized intersections. Using an observational (simple) before-and-after crash analysis, researchers found that two sites in which fatal crashes were found in the before condition observed no fatal crashes after the installation of a roundabout. No sites experienced a fatal crash after a roundabout was installed. For all injury (A, B, and C) crashes, the number of locations with a reduction in these crash types were greater than the number of locations with increases in these crash types. Similarly, the magnitude of decrease in injury crashes is higher than the magnitude of increase. When examining crash type, researchers found that roundabouts changed the crash types that occur at intersections from more severe crash types like angle and head-on to less severe crash types such as sideswipe and single vehicle crashes.