Abstract
A general human behavior feedback formalism is developed in which the actual safety change in traffic systems is related to the intended or expected change through the introduction of a "human behavior feedback parameter," f. This formalism includes earlier approaches to modeling or understanding traffic safety as special cases and in addition includes responses outside the range encompassed by the earlier approaches. It is applied in an essentially symmetric way to changes expected to increase or to reduce safety. Compared with the expected safety change, the actual safety changes found in 26 studies in the literature provide examples of all of the following: change even greater than expected; as expected; smaller than expected; no safety change; and a perverse effect (safety change opposite in sign to expected). It is concluded that human behavior feedback is a pervasive phenomenon in traffic systems, and one that may greatly influence the outcome of safety measures.

This publication has 51 references indexed in Scilit: