Measurements of vehicle vibration and passenger discomfort have been used to evaluate alternative methods of predicting passenger vibration discomfort. Twelve vibration inputs to the body were considered: three translational and three rotational axes on the seat, three translational axes at the back and three translational axes at the feet. Equivalent comfort contours were obtained by laboratory experimentation for all twelve inputs for eight male subjects. The responses of these subjects to the ride in six different cars on twelve different roads were then correlated with nine alternative methods of averaging the twelve vibration inputs.Evaluation of only the ‘worst’ vibration input or the ‘worst’ vibration frequency did not provide good predictions. Either the rms or rmq of the frequency and axis weighted accelerations should be determined. The square root of the sums of squares of these twelve values provided the best correlation with the discomfort ratings in the vehicles.