In known devices for testing vehicle shock absorbers, the tires, which have a dampening effect, are not taken into account, so that the measurement results may be falsified by up to about 30%, depending on the tire pressure. In addition, when the wheels are mounted on the vehicle, no shock absorber characteristic curve may be measured, only individual values. In the disclosed device, the actuating variable “tire” is eliminated by a regulation process that consists in normalizing the resonance amplitude, a considerably more accurate measurement being thus obtained. In addition, the spring constant of the wheel suspension is determined before actually testing the shock absorbers, allowing defective springs to be detected. Finally, car body oscillations may also be determined, causing an additional precision gain. For the first time it becomes possible to determine a shock absorber characteristic curve without having to dismount the shock absorbers.