This paper reviews experiments for determining human response to vibration and interprets them in the vehicle vibration context. It reviews the author's research and compares it with findings of other investigators. From the results of vehicle tests, it is concluded that properly weighted “acceleration variance” is a meaningful measure of vehicle vibration. For a single vehicle subjected only to surface profile changes, a single number summed over all frequencies can be used. For two or more vehicles with different vibrational characteristics, acceleration variance must be computed in several frequency bands to yield correction factors that define the nature of the vibration.