Perceptual estimates of action-relevant space have been reported to vary dependent on postural stability and concomitant changes in arousal. These findings contribute to current theories proposing that perception may be embodied. However, systematic manipulations to postural stability have not been tested, and a causal relationship between postural stability and perceptual estimates remains to be proven. We manipulated postural stability by asking participants to stand in three differently stable postures on a force plate measuring postural sway. Participants looked at and imagined traversing wooden beams of different widths and then provided perceptual estimates of the beams' widths. They also rated their level of arousal. Manipulation checks revealed that the different postures resulted in systematic differences in body sway. This systematic variation in postural stability was accompanied by significant differences in self-reported arousal. Yet, despite systematic differences in postural stability and levels of arousal perceptual estimates of the beams' widths remained invariant.