The chest impact data of Kroell, et al., and Stalnaker, et al., were examined to determine the relationships that might exist between the physical characteristics of cadavers, impact conditions, and responses. It was found that while the Kroell male, Kroell female, and Stalnaker data had similar physical characteristics, their responses were not related to their physical characteristics and impact conditions in the same manner. Regression equations were found that fit the Kroell male data extremely well. Based on a regression analysis of the Kroell male data, scaling rules were developed that allowed performance requirements for chest response to be defined for 5th, 50th, and 95th percentile dummies. Previous response requirements were for 50th percentile dummies only and were based on averaged responses from subjects whose average characteristics differed widely from 50th percentile characteristics.