METHODS FOR MEASURING EXPOSURE TO AUTOMOBILE ACCIDENTS

Thorpe's method for determining exposure to risk of accident by using information routinely available from most accident reporting systems is discussed, and a study is reported which compares a measure of induced exposure, based on reports from North Carolina (1971), with a measure of reported driving exposure provided by a state-wide sample of driver license applicants. Only applicants and drivers age 20 and over were included in tha analyses. Accident drivers were classified by age and sex; in-state versus out-of-state; single vehicle, guilty double, or innocent double; and vehicle make and year. Figures are presented which compare exposure distribution of drivers found guilty in two vehicle accidents; compare between exposure and drivers not in violation in two vehicle accidents, and compare the single vehicle accident drivers with the exposure distribution. The figures are discussed and conclusions reached are presented. Defensive driving appears to occur at a later age than avoidance of accident causation. Since the measure of exposure (adopted here) consists of driver's estimates of mileage driven, a partial check was made by comparison with an independent measure of summer exposure, namely trips included in an origin and destination survey. The differences between the 0 and D exposure measure and the earlier exposure measure are listed. The study indicates the impossibility of disproving the induced exposure concept although less confidence is expressed in the idea. If induced exposure is to prove a valuable tool, the issue of verification is of paramount importance.