A study on the effect of road remedial work on road accident frequency

Reducing road accident frequency is an endeavor of national importance. The engineering approach, particularly road remedial work, is often the preferred method due to its ease of implementation and cost-effectiveness. However, the limitation of this 'blanket approach' is that accidents often have specific causes, making blanket remediation potentially inappropriate. Due to its widespread use, verifying the effectiveness of remedial work is necessary, requiring the ability to perform an objective assessment (road auditing). This paper studies the use of the Poisson probability distribution as a tool to objectively measure the impact of road remedial work, based on the assumption that accident occurrences conform to a Poisson process. It is suggested that the Poisson distribution can approximate future accident frequency at a selected site and measure the effectiveness of remedial work. Research found that the accident process for the selected sites does conform to the Poisson distribution, indicating its appropriateness as a measurement tool. However, tests using the Poisson distribution on the selected sites found that the road remedial works undertaken were not able to reduce the mean accident frequency. The study concludes that the use of the Poisson distribution is appropriate and should be used as a basis to measure the effect of road remedial work.