In crash testing automobiles to investigate safety performance, many sizes of anthropomorphic dummies are employed, from the large 95th percentile males down to 6-month-old infant models. The on-board dummies play an important role in the overall crash test experiment because they simulate the human occupants and provide information on potential injury levels and kinematics of the passengers. In connection with the work of continually upgrading the anthropomorphic dummies to better simulate human occupants in crash work, this program was focused on the six-year-old child dummy.When employing a test dummy in typical car crashes, the performance repeatability of the unit must be known in order to obtain reliable and useful results. In this paper, the model SA106C six-year-old size dummy was subjected to numerous static and dynamic tests in an investigation of its performance repeatability. Standard calibration tests and accelerator sled runs were performed on four dummies using standard testing procedures and modern child restraints. The results were analyzed statistically which indicated, in general, that these model dummies were repeatable in performance.