Abstract
Involvement rate is commonly used as a measure of risk of being involved in an accident. Whenever involvement rate is expressed as a function of the deviation of the speed of the accident-involved vehicle from mean traffic speed, errors of measurement become extremely important. The effects of measurement errors, in both the numerator and denominator of the involvement rate, have been examined. Examples are given to illustrate, numerically, the effects of measurement errors. Some possible alternatives for reducing or eliminating these effects are presented.