Three experiments were conducted to compare the independent component method and the weighting method, recommended by current vibration standards for evaluating complex vibrations. Seated subjects matched their perceptions of the intensity of various Z-axis vibrations by adjusting the intensity of a sinusoidal matching frequency. Exp. I stimuli were composed of from one to four sinusoids in the frequency range from 11--63 Hz, Exp. II stimuli were made up of from one to four third-octave bands of random vibration with center frequencies from 16--40 Hz, and Exp. III stimuli were composed of from one to four sinusoids from 2.6--16 Hz. In all three experiments, the acceleration of the matching response increased significantly as the number of sinusoids or third-octave bands in the stimulus increased. Such a relationship is not predicted by the independent component method, but is predicted by the weighting method. Moreover, when the weighted accelerations of the matching responses are compared with the weighted accelerations of the input vibrations, the correspondence is quite good for all three experiments. These results support adoption of the weighting method as the preferred procedure for evaluating complex vibration environments.