| Traditionally, the measures of exposure fall into two general categories: direct and indirect. The former includes vehicle miles traveled (VMT), number of entering vehicles (NEV) for an intersection, and annual daily traffic (ADT). Indirect exposure is normally referred to as induced exposure. VMT is the most commonly used exposure measure in traffic safety/crash related analyses. However, VMT as an exposure measurement has also aroused criticism among traffic researchers: the underlying assumptions of VMT-based methods have been challenged. In addition, the use of VMT confronts two fundamental drawbacks in practical applications: general availability of data and finer disaggregation of exposure.; Quasi-induced exposure, an approach to estimate relative exposure, is capable of overcoming those difficulties confronted by the VMT method. Quasi-induced exposure has been employed by a number of traffic researchers and has demonstrated its strength in a variety of practical applications. For example, it is relatively easy to use; incorporated accident data are generally available; it is able to measure the exposure of a specific driver cohort under specific conditions (e.g., young male on Friday night on local streets). However, the theory of quasi-induced exposure is far from being perfect. A set of issues needs to be addressed: (1) there is a lack of a systematic procedure to prepare accident data; (2) there are problems involved with the responsibility assigning scheme; (3) more importantly, there are few attempts to validate the underlying assumption: non-responsible drivers in two-vehicle accidents constitute a random sample of driving population on the road.; Two techniques have been developed to validate the underlying assumption: (1) to compare the relative exposure derived from accident data using quasi-induced exposure with the "true" exposure collected from other sources, such as VMT data, seat belt use data, and truck volume data; (2) to compare the distributions for non-responsible drivers derived from two-vehicle and three-or-more-vehicle accidents.; By means of addressing these three stated issues, this research aims to develop a guideline with regard to when to, or not to, use the quasi-induced exposure technique, and to provide a systematic procedure to manipulate accident data and assign accident faults if the validity of quasi-induced exposure is ensured. |
