Safety Evaluation of Wet-Reflective Pavement Markings

The Federal Highway Administration organized a pooled fund study of 38 States to evaluate low-cost safety strategies as part of its strategic highway safety effort. One of the strategies selected for evaluation was the application of wet-reflective pavement markings. This strategy involves upgrading existing markings from standard marking materials to wet-reflective markings applied as a paint, tape, or thermoplastic material. The purpose was to provide an improved level of retroreflectivity in wet-road conditions. Geometric, traffic, and crash data were obtained for treated freeway sections in Minnesota, North Carolina, and Wisconsin; treated two-lane rural road locations in Minnesota; and treated multilane road sections in Wisconsin. To account for potential selection bias owing to regression-to-the-mean, an Empirical Bayes (EB) before–after analysis was conducted. The analysis also controlled for changes in traffic volumes over time and time trends in crash counts unrelated to the treatment. Intersection-related, snow/slush ice, and animal crashes were excluded from the analysis. For freeways, the combined results for all States indicated reductions in crashes that are statistically significant at the 95-percent confidence level for injury and wet-road crashes, with estimated crash modification factors (CMFs) of 0.881 and 0.861, respectively. For multilane roads, statistically significant reductions were estimated for total crashes (CMF = 0.825), injury crashes (CMF = 0.595), run-off-road crashes (CMF = 0.538), wet-road crashes (CMF = 0.751), and nighttime crashes (CMF = 0.696). For two-lane roads, the sample of crashes was too small to detect an effect with statistical significance for any of the crash types, but there were indications that the treatment had a safety benefit for wet-road crashes. Benefit–cost ratios estimated with conservative cost and service life assumptions were 1.45 for freeways and 5.44 for multilane roads. The results suggest that the treatment—even with conservative assumptions on cost, service life, and value of a statistical life—can be cost effective, especially for multilane roads.