A hybrid mathematical model of a prototype baseline vehicle including dummies and side impact barriers were used to evaluate the effect of vehicle modifications on injury response. Models of the EUROSID-1 and US-SID dummies, and a model of the human body were included in the vehicle model. Models of the EEVC foam and NHTSA honeycomb side impact barriers were used as impacting object. The vehicle model was impacted at 15 mph (24 km/h), 25 mph (40 km/h) and 35 mph (56 km/h) by the barriers. The occupant responses of vehicle modifications were evaluated.At 15 mph (24 km/h) barrier impact velocity, no reductions in the injury criteria were obtained with padding. At 25 mph (40 km/h) and 35 mph (56 km/h) barrier impact velocity, however, reductions in the injury criteria were obtained with padding. The impact with the EEVC barrier was found to be more violent for the occupant than the impact with the NHTSA barrier; thus, higher dummy responses were obtained with the EEVC barrier. It was shown that vehicle improvement to reduce injury criteria in side impacts did not necessarily entail the stiffening or strengthening of the structure. It was found that modifying the structure to minimize the dummy responses for one side-impact barrier may at other impact velocities and other impacting objects result in increased injury criteria.