The accuracy of determining speeding directly from mass crash data and using the NSW Centre for Road Safety method

Abstract

Exceeding the posted speed limit, or speeding, is generally accepted as a major cause of road crashes and in particular fatal crashes. However, the actual proportion of crashes in which one or more vehicles was speeding is not easily determined. The exact travelling speed of a vehicle prior to a crash can only be determined by detailed crash reconstruction. Such a reconstruction is considered beyond the scope of regular traffic police who record the majority of the crash data that makes up the mass crash databases such as the South Australian Traffic Accident Reporting System (TARS). It is therefore thought that speeding is underreported in the mass crash data. A method was developed by NSW to identify, from mass data, crashes that involved speeding as a factor. This method was subsequently used by other states, including South Australia. The Centre for Automotive Safety Research conducts the crash reconstructions required to determine speed as part of its at-scene in-depth crash investigation work. This paper compares the actual proportion of speeding crashes in the most recent set of at-scene in-depth crash investigation cases with that found by using the mass data and the method developed by the NSW Centre for Road Safety. It was found that the error ‘excessive speed’ recorded in the TARS database is not accurate in identifying crashes where a vehicle was speeding. The NSW Centre for Road Safety method of determining speeding in crashes was also found to lack accuracy, though it was more accurate than simply relying on the error ‘excessive speed’ in the TARS database.