Due to the high correlation that generally exists among driver characteristics, as well as the overall complexity of the factors involved, the role that gender plays in the relationship between drivercharacteristics and accident risk has been difficult to quantify using traditional statistical approaches. This paper attempts to provide some new insight by using hazard functions and a sample of University of Washington drivers. The subsequent empirical analysis uncovers important differences in the relationship between male and femaledriver characteristics and their respective accident risk. The application of the hazard function approach is achieved by recording theinterval between starting driving and being involved in an accident, and the time between any subsequent accidents, and then statistically determining the effect that driver characteristics have on this.The data was obtained from a questionnaire filled in correctly by 114 male and 46 female drivers. The estimation results of the Weibullproportional hazards model of the duration of time until accident occurrence broken down by: driving experience, whether the driver is married or not, accident record, and household income, are presentedand discussed. These results show that male and female drivers havesignificantly different driver characteristic/accident risk relationships. In both sexes, the risk of having an accident increases withnumber of miles driven per year. Single people and those on lower incomes have lower accident risk. The longer a male driver goes without having an accident, the less likely he is to have one. With women, the likelihood of having an accident is independent of the time that has transpired without having an accident
Abstract