Sensible management of traffic safety is based on having reasonable expectations about the effect of various safety programmes and countermeasures. unfortunately, in spite of decades of research and experience, the safety effect of many countermeasures is largely unknown. this sorry state of affairs is mainly due to the fact that it is difficult to conduct conclusive experiments about the effect of safety countermeasures. recognition of this objective difficulty should lead one to question the usefulness of classical statistical tests of significance as a device for scientific progress in this field. the unquestioning and all-pervasive use of significance testing in evaluative research on safety amounts to a self-inflicted learning disability. the first part of the paper explains why. the fundamental problem of evaluative research is how to build knowledge from experimental results. recognizing that in transport safety knowledge is gradually accumulated from the results of many studies, alternative approaches are explored. the application of a Bayesian approach in a specific case is described.(a) for the covering abstract of the conference see IRRD abstract no 264967.
Abstract