An analysis is presented that was used to interpret the significance of response measurements made with a specially instrumented, 3-year-old child dummy that was used to evaluate child injury potential of the second-generation, passenger inflatable restraint system that was being developed by General Motors Corporation. Anesthetized animals and a specially instrumented child dummy, both 3-year-old child surrogates, were exposed to similar inflating-cushion, simulated collision environments. The exposure environments were chosen to produce a wide spectrum of animal injury types and severities, and a corresponding broad range of child dummy responses. For a given exposure environment, the animal injury severity ratings for the head, neck, thorax and abdomen are paired with dummy response values corresponding to these body regions. These data are used to develop relationships that can be used to predict the probability of an animal experiencing significant injuries to these body regions based on the child dummy response measurements. A rationale is developed for interpreting the predicted animal injury severities relative to child injury severities.