Dupont E., Martensen H., Stipdonk H., Bijleveld F., Commandeur J., Papadimitriou E., Yannis G., Antoniou C., Bergel R., Brandstaetter C., “Analyses using the European Road Safety Observatory”, Proceedings of the 4th IRTAD Conference, Seoul, September 2009.

Road safety analyses aim to describe and explain road safety outcomes (road accidents and casualties), either in time or in space, as well as to forecast future developments on the basis of existing experience. The availability of reliable data is one of the most important conditions for useful road safety analyses. Within the development of the European Road Safety Observatory (ERSO), a wealth of road safety related data at European level was gathered, harmonized and made available, including road accidents data (through the CARE database), risk exposure data, road safety performance indicators and in-depth road accident investigation data. The objective of this paper is to present a comprehensive overview of analyses carried out by using data and knowledge from the European Road Safety Observatory. These analyses were carried out within the framework of the EC co-funded integrated research project SafetyNet and can be broadly distinguished into time series analyses, geographical analyses, and accident analyses. It is shown that different types of dependencies are present among observations (e.g. serial or hierarchical dependencies), calling for particular statistical treatment by means of dedicated techniques, such as multilevel analysis and time series analysis. Moreover, additional methodological questions are tackled, such as the use of exposure data; the interest of disaggregate analyses; and the potential of combining ERSO data with other data sources. A brief overview of some interesting findings from related case studies in different countries or at European level is finally provided, including combined analyses of road accidents, fatalities and exposure, structural and explanatory analyses of road safety trends, disaggregate time series analyses and forecasts, spatial modeling, joint analysis of CARE and road safety attitude data, statistical analysis of injury underreporting rates and modeling of in-depth fatal accident investigation data.