Abstract
This paper examines how road-user behavior changes in response to changes in traffic system safety thereby generating surprising outcomes from safety interventions. In response to changes introduced to increase safety, examples are presented in which the actual outcome was that: (1) safety increased even more than expected; (2) safety increased as expected; (3) safety increased, but less than expected; (4) no observed change in safety; (5) the change actually decreased safety (a perverse effect). Correspondingly diverse reactions were found for changes expected to reduce safety, but introduced to achieve other goals. (Author/publisher).