Biomechanical headforms are used for helmet certification testing and reconstructing helmeted head impacts;​ however, their biofidelity and direct applicability to human head and helmet responses remain unclear. In addition, proper helmet fit is important for optimizing head protection during an impact, yet many motorcyclists wear helmets that do not properly fit their heads. In the first phase of this research, the head dynamics and helmet foam liner deformations generated by cadaver heads and three headforms were compared during motorcycle helmet impacts. In the second phase of this research, the effect of a mismatch in headform and helmet size on the headform response and helmet residual foam liner deformation was examined. In Experiment 1, four cadaver heads and three headforms (50th percentile Hybrid III, International Standards Organization (ISO), and Department of Transportation (DOT)) wearing a shorty-style motorcycle helmet were dropped onto the forehead region against a flat anvil using initial kinetic energies of 75, 150, and 195 J (impact speeds between 5.4 and 9.3 m/s). Computed tomography (CT) scans were used to quantify the maximum residual crush depth and residual crush volume of the helmet foam liners. General linear models were used to quantify the effect of head type and impact energy on peak linear head acceleration, head injury criterion (HIC), impact force, maximum residual liner crush depth, and residual liner crush volume. Linear regression models were then used to separately quantify the relationship between the response variables (peak acceleration and impact speed) and the predictor variables (maximum crush depth and crush volume). The cadaver heads generated larger peak accelerations than all three headforms, larger HICs than the ISO, larger forces than the Hybrid III and ISO, larger maximum crush depth than the ISO, and larger crush volumes than the DOT. These significant differences between the cadaver heads and headforms showed that none of the headforms replicated all of the biomechanics of the cadaver heads, and that these differences need to be considered when attempting to estimate an impact exposure using a helmet’s residual crush depth or volume. In Experiment 2, four sizes of a shorty-style helmet were tested on four sizes of ISO headforms during forehead impacts against a flat anvil using initial kinetic energies between 10 and 275 J (impact speed between 2.0 and 10.5 m/s). CT scans were again used to quantify maximum residual crush depth and residual crush volume of the helmet foam liners. Separate linear regression models were used to quantify how the response variables (peak acceleration, HIC, and impact speed) were related to the predictor variables (maximum crush depth, crush volume, and the difference in circumference between the helmet and headform). The results indicated that increasingly oversized helmets reduced peak headform acceleration and HIC for a given impact speed. Peak headform acceleration, HIC, and impact speed can be estimated from a helmet’s residual crush for maximum residual crush depths less than 7.9 mm (31-34% of original foam thickness) and residual crush volumes less than 40 cm3 . Above these levels of residual crush, large variations in headform kinematics are present, possibly related to densification of the foam liner during the impact. Overall, this research has shown that helmet impact kinematics can be estimated from residual helmet liner deformation under some conditions, although differences between the human head and biomechanical surrogates for the human head need to be considered when making these estimates.