Potential crash reduction benefits of shoulder rumble strips in two-lane rural highways

  • Published: Accident, Analysis & Prevention, 75, 35-42, 2015
  • Authors: Khan, M., Abdel-Rahim, A & Williams, C.J
  • Date Added: 05 Apr 2016
  • Last Update: 05 Apr 2016
  • Format: pdf


Examine the effectiveness of shoulder rumble strips in reducing run-off road crashes on rural highways.


Empirical Bayes before and after analysis was used to measure the impact of installing shoulder rumble strips on highways in Idaho. The crash data were obtained from four different sources (vehicle crash report data from Idaho’s Office of Highway Safety; the office of highway operation and safety; yearly vehicle exposure data from the automatic traffic recorders; and satellite images from Google Earth for roadway sections) and covered the period between 2001 and 2009. Rumble strips were installed on three highways between 2004 and 2007, providing 38 treatment sites (total length of 178.63 miles).

Key Findings:

  • The results showed a 14% reduction in all run-off-road crashes after the installation of the rumble strips.

  • In addition, the total number of run-off-road crashes on test sites was 92, compared with 106.5 on control sites.

  • The road geometry also had an impact, with rumble strips being most effective on roads with a moderate horizontal curve where there was a 29% reduction in run-off-road crashes (compared to 22% on roads with no horizontal curves and 8% on roads with sharp horizontal curves).

  • The width of the right hand shoulder also had an impact, with the most significant reduction being found for right paved shoulders measuring 3 feet or more


Shoulder rumble strips, run-off-road crashes


Empirical Bayes analysis allows for a before and after comparison, while accounting for the regression to the mean effect and external causal factors that change with time. It also considers trends at the treatment sites prior to the installation and the safety performance and crash trend at similar control sites.