Vestibular stimulation in the form of rocking movements could be a promising non-pharmacological intervention for populations with reduced sleep quality, such as the elderly. We hypothesized that rocking movements influence sleep by promoting comfort. We assessed whether gentle rocking movements can facilitate the transition from wake to sleep, increase sleep spindle density and promote deep sleep in elderly people. We assessed self-reported comfort using a pilot protocol including translational movements and movements along a pendulum trajectory with peak linear accelerations between 0.10 and 0.20 m/s(2). We provided whole-night stimulation using the settings rated most comfortable during the pilot study (movements along a pendulum trajectory with peak linear acceleration of 0.15 m/s(2)). Sleep measures (polysomnography) of two baseline and two movement nights were compared. In our sample (n = 19; eight female; mean age: 66.7 years, standard deviation: 3 years), vestibular stimulation using preferred stimulation settings did not improve sleep. A reduction of delta power was observed, suggesting reduced sleep depth during rocking movements. Sleep fragmentation was similar in both conditions. We did not observe a sleep-promoting effect using settings optimized to be comfortable. This finding could imply that comfort is not the underlying mechanism. At frequencies below 0.3 Hz, the otoliths cannot distinguish tilt from translation. Translational movement trajectories, such as used in previous studies reporting positive effects of rocking, could have caused sensory confusion due to a mismatch between vestibular and other sensory information. We propose that this sensory confusion might be essential to the sleep-promoting effect of rocking movements described in other studies.