Blackspot remediation and options for improving the process

Abstract

A blackspot refers to a single site that has an unusually large number of accidents --- whether for obvious reasons such as a great deal of traffic, or for non-obvious reasons that become apparent only after careful investigation. Part 1 of this paper (TEC, October 2011) drew attention to the small number of crashes that (in developed countries) typically constitute the criterion for identifying a blackspot. At that number of crashes, the actual number inevitably varies a great deal from year to year (or from one three-year period to another), and the selection of sites by this means is consequently unreliable. This paper will first examine the busiest blackspots: how much might it be worth spending at busy blackspots, and how severe are accidents at busy blackspots. Then some possible ways in which the blackspot remediation process might be improved will be discussed. (The empirical Bayes method. Randomised trials. Should the evidence from the crash record and from the site inspection be more independent? Crash debris and traffic conflicts. Administrative processes. Geographical databases as a source of characteristics of crash sites.) Finally, there will be concluding discussion. A greater emphasis on expert judgment, based largely on site inspection, might be an improvement. Blackspot programs are believed to be highly cost-effective, and this is not brought into issue: the question concerns whether even better results are likely to be obtained by a procedure that is less dependent on the vagaries of crash numbers.