The paper mainly deals with methodological foundations of crashworthiness rating procedures based on real world accident material. Most rating methods, especially those for car-to-car crashes, are based on so-called contingency tables containing the injury outcomes for the drivers of the subject and the opponent car. Some light is shed on the assumptions usually made and it is argued that data material, which contains information on the injury severity, is much better suited for rating procedures.

Additionally the paper demonstrates for characteristic real data examples that there exist essential factors, which influence rating results in a non-negligible way.

Finally, the strong relationship between accident severity and injury outcome is investigated.

All theoretical results are underpinned by real accident data taken from the Hanover and Dresden data collected within the German In Depth Accident Study (GIDAS).