Abstract:
Given that more and more indicators are used to describe the complex character of the road safety phenomenon, the development of a road safety index is increasingly recognized as a useful tool to reduce the size of the topic without dropping underlying information and to perform meaningful comparison and monitoring of overall road safety performance. In this respect, a scientifically sound methodology for index construction is crucial. In the SUNflowerNext study, a comprehensive road safety index that aggregated 21 sub-indicators belonging to four different types of road safety indicators was created for a set of European countries. As a revisit of this study, this paper focuses on two methodological challenges when developing such an index. One is to reflect the hierarchical structure of the indicators, and the other is to distinguish between quantitative and qualitative data. To this end, a mathematical programming model based on the data envelopment analysis approach is proposed. By solving a constrained optimization problem, the concept of layered hierarchy is embodied in the model and the presence of both quantitative and qualitative data is properly integrated. Based on the country-specific models, the optimal road safety index score is computed for each country, and a more acceptable country ranking is obtained compared with the one from the SUNflowerNext study. In addition, by deducing the weights allocated in each layer of the hierarchy for each country separately, valuable information on prioritizing policy action per country is derived.
