Researchers worldwide try to define a unique test procedure for the assessment of whiplash protection of seats and restraint systems in low speed rear-end impact. Apart from valid injury criteria and uniform crash conditions, there is no clear answer to the question, which dummy to use. There are two impact dummies currently available, which have been designed for rear-end impact testing: BioRID and RID2. Both dummies have been evaluated in several test programs, however, both dummies have never been compared with each other in the test conditions, which form the basis of their design. BioRID was based on and validated against volunteer tests performed by Davidsson and Ono, while RID2 was designed with and validated against PMHS tests done by Bertholon and compared to volunteer tests reported by Van den Kroonenberg. This paper compares the responses of both rear impact dummies and the Hybrid III for the test conditions mentioned above. The setup of Davidsson used a rigid seat with flexible back and head restraint panels, while the setups from Ono and Bertholon used a rigid seat without a head restraint, in spite of being not representative for real car seats. This configuration creates a well defined test environment which will not affect nor obscure the dummy response Results of the performance of both rear impact dummies and the Hybrid III in comparison to the human responses will be presented in this paper. The results show that both rear impact dummies are capable of simulating rear impact responses, especially the head-neck kinematics. A difference in load pattern was found, which could be relevant when injury criteria will be based on neck forces and/or torques. Moreover, the dummies show a different interaction with the seat back, illustrated by the differences in T1 kinematics: BioRID shows larger T1 rotation and more ramping up than RID2, while spine straightening is comparable for both dummies. The current study showed good scores for both dummies in the setup on which they are based. The biofidelity score of BioRID is slightly better than for RID2, while the performance of the Hybrid III is relatively poor. However, repeatability, reproducibility and handling are not part of the evaluation, even though they are important for the practical use of the dummies.