The paper describes a sound methodology, which leads to a GIDAS-aided quantification of the effectiveness of traffic safety measures for passenger vehicles for countries in EU 27.
Even based on rather limited accident information from national statistics and under the assumption that comparable accidents in different countries lead to comparable accident outcomes, the described procedure allows defining so-called weighting factors. By weighting each single accident of GIDAS a modified GIDAS database can be established which imitates the accident situation in the region or country of interest to some extend. The main ingredients of the proposal are a proper clustering of European countries according to their accident occurence and a statistical procedure (Iterated Proportional Fitting) which allows the prediction of the common distribution (high dimensional) of accident data of the region or country of interest based on available lower dimensional marginal distributions (even one-dimensional). Since the effectiveness of safety systems reliably can be predicted on the detailed GIDAS accident database our procedure allows a prediction of the effectiveness of such systems for other regions or countries. The method is confirmed by real accident data examples, which shed light on the strengths and weaknesses of the proposed procedure.