The before and after technique was used to evaluate 37 widening projects representing 143 miles (230 km) of improved road. Widening was essentially completed on existing alignment. Accident rates were reduced for all three shoulder widening widths of 28 (8.5 m), 32 (9.8 m) and 40 feet (12.2 m). The accident rate was reduced 16 percent for 28-feet under 3,000 AADT, 35 percent for 32 feet under 5,000 AADT and 29 percent for 40 feet over 5,000. The reductions were statistically significant for 32 and 40 feet at the 95 percent confidence level. The study of 23 projects with 51 miles (82 km) showed that accidnet rates were reduced for 2-lane roads with a passing lane. The rates were reduced 11 percent for 36 feet (11.0 m) 25 percent for 40 feet (12.2 m) and 27 percent for 42 to 44 feet (12.8 m to 13.9 m). The reductions were significant for 40 and 42 to 44 feet. Forty feet is considered the minimum adequate width for passing lanes. Accident reports show that the one-way barrier stripe (single lane direction permitted to pass) may cause problems. This stripe may create a false sense of security for the drivers. Drivers appear to move into the center passing lane without looking. The one-way barrier stripe should be used with extreme care to prevent the old 3-lane highway situation of severe head-on collisions.
Abstract