North Carolina data are used to illustrate some points concerning the pros and cons of quantifying state highway safety goals. It is recommended that impact goals be left unquantified rather than be measured by committee vote or some other nonobjective procedure. In the highway safety field, there is frequently little objective evidence for determining the presence or amount of a program's impact on accidents. Three approaches to setting program goals are outlined, including their limitations: use of previous evaluations, "break even" cost analysis, and statistical detection of change. It is emphasised that the National Highway Traffic Safety Administration and the states should attempt to develop better goal-setting procedures if such quantified goals are to remain an integral part of the Highway Safety Program. Much greater use of project impact evaluation is strongly advocated.
Abstract