Generalised linear models of road trauma outcomes have been found to be a powerful way of representing the trends and variations over time and to explain the effects of influential factors such as countermeasure initiatives. This report covers their application to monthly casualty crash frequencies and injury severity outcomes in Victoria during 1998 to 2003. During 2000 to 2002, the mobile speed camera program in Victoria was changed by introducing “flashless” camera operations during daytime and other modifications to make the enforcement more covert and unpredictable, increasing the targeted camera operating hours from 4200 to 6000 hours per month, and reducing the speeding offence detection threshold in three stages. In December 2002, the thresholds for penalties applying to different levels of speeding offence were generally reduced by 5 km/h. Associated with the speed enforcement initiatives was a program of speed-related advertising known as the “Wipe Off 5” campaign launched in early August 2001 without specific reference to the enforcement changes. A subsequent announcement took place in late November 2001 specifically mentioning the more covert speed camera operations as well as the increase in camera hours. A third announcement took place in the print media at the end of March 2002, emphasising that the former 9 km/h speeding tolerance no longer applied. As well as the enforcement changes, the statistical models included the reduced urban speed limit in January 2001, the penalty restructure, the speed-related advertising and the announcements, plus the impact of the fixed speed camera controversy that arose in late 2003. The study concluded that generalised linear modelling of crash outcomes as a function of potential explanatory factors needs realistic assumptions to be made about viable functional forms connecting a measure of each factor and the outcomes. The assumed functional form of the relationship between monthly speed camera hours and road trauma appears to represent this relationship well. There is doubt that the flashless speed camera initiative and the enforcement threshold reductions have been adequately represented in the monthly crash outcome models. It was concluded that the effect of these initiatives on crash outcomes is unknown at this stage. The assumed functional form of the relationship between monthly speed-related advertising levels and road trauma appears to represent this relationship well. The speed-related television advertising had a statistically significant association with a decrease in monthly casualty crash frequencies during times of increased advertising levels. In general the relationships connecting speed camera hours and levels of speed-related advertising with road trauma reductions in Victoria confirmed previous research on the effectiveness of these road safety programs as operated in the State. (Author/publisher)
Samenvatting