Frontal sled tests of the Part 572, APR, and Hybrid III dummies were conducted in a three-point restraint system at 50 km/hr velocity change. The tests were conducted to evaluate the dummy responses in a tightly controlled systems environment, and to compare the dummy responses to previously established cadaver responses from the same environment. The Hybrid III dummy measurement repeatability was found to be better than either the Part 572 or APR dummies, although the thoracic acceleration responses from all three are shown to be quite similar to cadavers. Correlation of the dummy measurements are made to a limited amount of both the cadaver data and accident data from the National Crash Severity Study.