Abstract
Considerable attention is still given to developing and using alternate methods for determining exposure for calculating highway accident rates. A quasi-induced method of measuring exposure developed in the late 1960s is re-examined and found to be promising for determining relative accident involvement rates. A new empirical investigation is offerred as the first step in verifying that the characteristics of the "innocent victim" in two-vehicle highway accidents represent a random sample of the driver-vehicle combinations present on the highway system under specified conditions. Quasi-induced exposure estimates are shown to be, at a minimum, consistent and reproducible.(A).