Traffic Conflict Approach in Road Safety: A Review of Data Collection Methods

Traditionally, safety analysis of road infrastructure has relied on the use of crash data accumulated at a specific location to understand crash risk and determine adequate measures to improve safety. However, crash data can be undermined by several limitations including long observation period required to collect sufficient data for reliable analysis, unavailability of data, lack of spatial and temporal precision, misclassification, underreporting, and post hoc assessment of crash causes. In addition, the reactive nature of crash-based assessment poses an ethical dilemma as one must wait for crash to occur, thus for people to suffer. An alternative and proactive approach consists of analysing road safety by studying non-crashes events called traffic conflicts. On the contrary of crash data which when available, often only require to be extracted from the different existing sources, traffic conflict data are usually almost not readily available for the situation of interest and need to be observed/recorded or generated. This paper proposes a review of the main data collections technique used in conflict-based road safety analysis studies. The methods are described, and qualitatively characterized on various dimensions including the types of indicators generated, the types of road facilities considered, and the application purposed, followed by a highlight of their strengths and limitations. The paper is expected to provide an overview of the potential opportunities that the different data collection methods present, and their suitability based on the research to be conducted.