Safety Evaluation of Cable Median Barriers in Combination With Rumble Strips on Divided Roads

CHAPTER 7. BEFORE–AFTER EVALUATION RESULTS

This chapter presents the results of the before–after evaluation, including aggregate and disaggregate analysis for Kentucky, Illinois, and Missouri.

Aggregate Analysis

Table 15 through table 26 present the resulting CMFs. These tables provide the estimates of predicted crashes in the after period without treatment, the observed crashes in the after period, and the estimated CMF and its SE for all crash types considered. CMFs statistically different from 1.0 at the 95-percent confidence level are indicated with an asterisk. The tables provide separate results for Illinois and Kentucky, the combined results for Illinois and Kentucky, and the results for Missouri. The research team did not combine the results from Missouri with the results from Illinois and Kentucky because the before conditions in Missouri did not include inside shoulder rumble strips, while the before conditions in Illinois and Kentucky included inside shoulder rumble strips.

Most of the Kentucky data were from rural roads, whereas most of the Illinois data were from urban roads. Because cross-median crashes are relatively rare, it would be difficult to draw any definitive conclusions unless a relatively large sample of sites were used. As a result, the focus should be on the combined rural and urban results. The combined urban and rural results from Kentucky indicated an increase in total crashes and a reduction in cross-median crashes. The results regarding total and cross-median crashes were similar in Illinois (for rural and urban combined), but the Illinois sites exhibited a reduction in injury and fatal crashes as well. The combined Illinois and Kentucky results (for rural and urban combined) indicate a 27-percent increase in total crashes, 22- to 24-percent decrease in injury and fatal crashes (depending on whether injury crashes were defined as KAB or KABC), and a 48-percent decrease in cross-median crashes. The increase in total crashes—along with a decrease in injury and fatal crashes—reveals an increase in property-damage-only (PDO) crashes. Based on the difference between the EB predicted crashes in the after period for total and injury and fatal crashes, the predicted PDO crashes in the after period (without the cable barrier treatment) were about 2,255.2 (3,319.37 minus 1,064.22), and the actual PDO crashes in the after period were 3,397 (4,208 minus 811). This implies that PDO crashes increased by approximately 51 percent following the implementation of the cable barriers.

The results from Missouri for total and injury and fatal crashes (for rural and urban combined) were very similar to the combined Illinois and Kentucky results. However, the reduction in cross-median crashes in Missouri was much more dramatic, with CMFs of 0.041 (based on cross-median indicator only) and 0.119 (based on cross-median indicator plus head-on).

Disaggregate Analysis

The disaggregate analysis sought to identify those conditions (i.e., before-period EB predicted crash frequency, median width, speed limit, and AADT) under which the treatment was most effective. The research team could discern no patterns, most likely because of the limited sample size for cross-median crashes.