A large number of experiments involving cadavers - including real-world-accident reconstructions - have been performed for the purpose of enhancing the state of knowledge concerning tolerance levels and protection criteria relevant to side-impact conditions. However, the scatter of the findings, as well as the considerable differences in injury severity levels (differences that cannot be accounted for by age differences alone) have limited the conclusions that it was possible to draw from these investigations in terms of criteria, mainly concerning thoracic protection.The major cause of scatter is the considerable differences in skeleton quality between subjects. Analysis of the rib characterization test findings made it possible to define a thoracic resistance index enabling the establishment of a classification of subjects. This index, which was validated with our sample, allowed us to evaluate the pertinence of the various side-impact protection criteria considered. When this index is used, thoracic deflection, in particular, emerges as a satisfactory criterion for predicting the risk of occurrence of rib fractures in side-impact collisions. The transposition of this criterion to dummies is consequently made possible, and is discussed.