Methodologies Used to Analyze Collision Experience Associated with Speed Limit Changes on Selected California Highways

Three methodologies are used and compared to determine whether there has been a statistically significant change in traffic collisions (total, fatal, fatal plus injury, wet pavement, and nighttime) due to recent speed limit increases on California state highways. The three methods compared were simple regression, analysis of variance (ANOVA), and an observational before–after study developed by Ezra Hauer. Both collision counts and rates were studied for early 1996 speed limit increases from 55 to 65 mph, and from 65 to 70 mph. A comparison group of highways that remained at 55 mph was also studied. The simple regression methodology did not detect nearly as many statistically significant increases as did the ANOVA and Hauer observational before–after methodologies. The latter two methods revealed very similar results, especially for statistically significant increases in total and fatal collision rates and counts. These methodologies also revealed a significant increase in nighttime collisions for the 55–65 mph group. The difference was that the observational method detected only a marginally significant increase with total collisions for the 65–70 mph group. Also, the ANOVA indicated that the fatality collision rate increase for the 65–70 mph group was marginally significant.