Validation data for child anthropomorphic test devices (ATDs) are scarce, making it difficult to assess their biofidelity. The goal of this study was to use previously collected real-world data involving a frontal crash with a child occupant to assess the biofidelity of current child dummies. The 9-year old child’s anthropometry placed him between the size of the Hybrid III 10-year old and 5th-percentile Adult Female (AF5) dummies. Though injuries on the child indicated that he was properly belted, there were contact points on the vehicle interior and the exact position of the child before the accident could not be definitively determined from the crash investigation. Sled tests with identical seat belts and bench seat were conducted with the HIII-10 year old (n=9) and AF5 (n=6) in various seating configurations to explore the possible posture of the child before the accident. The tests were designed to reproduce the predicted Delta-V of 51 km/h with a smaller subset of the tests performed at 59 km/h to assess the implications of a higher speed on occupant contacts. Video analysis was performed to determine trajectories of the dummy head, chest, pelvis, and extremities. Despite the variation in speed, neither dummy was able to achieve the maximum head excursion necessary to make contact with the dash board. The results suggest that the dummies may underestimate the magnitude of excursion experienced by the child involved in the actual crash. To further investigate this finding, a sensitivity study was carried out using MADYMO Hybrid III 5th percentile female model. In addition to making use of existing data to further the investigation of child dummies, this study examines the biofidelity of two dummies used in child response approximation.