The impacts of speed cameras on road accidents: an application of propensity score matching methods

  • Published: Accident Analysis & Prevention, Volume 60, November 2013,
  • Authors: H. Li, D.J. Graham, A. Majumdar
  • Date Added: 11 Feb 2016
  • Last Update: 11 Feb 2016
  • Format: pdf


To evaluate the impacts of speed limit enforcement cameras on reducing road collisions in the UK by accounting for both confounding factors and the selection of proper reference groups.


Using the propensity score matching (PSM) to evaluate the impacts of speed cameras, compared to using the empirical Bayes (EB) method and a naïve before and after approach.

Key Findings:

  • A total of 771 treatment sites and 4,787 potential reference group sites were observed for a period of 9 years in England. 

  • Confounding factors were accounted for using the propensity score matching (PSM), compared to empirical Bayes (EB) and before and after analysis.

  • Both the PSM and EB methods show similar results that there are significant reductions in the number of collisions of all severities at speed camera sites.

  • It is suggested that the propensity score can be used as the criteria for selecting the reference group in before-after control studies.

  • Speed cameras were found to be most effective in reducing collisions up to 200 metres from camera sites and no evidence of collision migration was found.

  • The analysis found no evidence of ‘kangaroo effect’ (i.e. no increase in collisions upstream and downstream camera sites). This is an important finding in that it shows that drivers do not alter their behaviour to deliberately decelerate and accelerate abruptly before and after the camera sites. Rather speed cameras have a constant effect on driver behaviour in reducing their speed.


Large database using different methodological approaches to assess camera effectiveness.