No abstract

No abstract

Unlike exogenous estimates of exposure to risk such as vehicle miles of travel, number of registered vehicles, and number of licensed drivers, quasi-induced (Q-I) exposure has not received adequate vetting. A criticism of Q-I is that its underlying assumptions are not convincingly validated or verified, partially because the risk estimates of Q-I have not been sufficiently compared with the more conventional techniques. The 2009 National Household Travel Survey data were used to derive annual vehicle miles traveled, disaggregated by characteristics of interest (age and gender). Comparisons were developed at different disaggregation levels between the vehicle miles traveled and the relative exposure calculated with Q-I. The main findings of the exercises follow: (a) statistical results suggest that the exposure estimates for 15 age groups and driver gender are in good agreement with the corresponding annual vehicle miles traveled and thus the induced exposure estimates are deemed to be reasonably representative of the driving population and (b) the validation study revealed that data disaggregation improves the homogeneity of age and gender distributions (reduced data irregularities caused by the aggregated distributions). The comparisons confirm that Q-I is a promising and powerful tool for estimating exposure in safety analysis.

No abstract

No abstract

No abstract

No abstract

The research illustrates that a series of environmental and crash factors (e.g., rear-end straight crashes, urban roadways, alcohol usage, and different driving cohorts) are statistically significant in interpreting the disparity of coefficients between 2- and 3-vehicle crash injury severity models. It raises awareness that the combined analysis of 2- and 3-vehicle crashes should be exercised with caution, particularly when safety research targets crashes with less severe injuries.