Evidence for an Influence of ‘Presence’ on Judgments of Egocentric Distance in Immersive Virtual Environments

This paper reports the results of a study intended to investigate the possibility that cognitive dissonance in ‘presence’ may play a role in the widely reported phenomenon of underestimation of egocentric distances in immersive virtual environments. In this study, we compare the accuracy of egocentric distance estimates, obtained via direct blind walking, across two cognitively different immersion conditions: one in which the presented virtual environment consists of a perfectly registered, high fidelity 3D model of the same space in which the user is physically located, and one in which the presented virtual environment is a high-fidelity 3D model of a different real space. In each space, we compare distance estimates obtained in the immersive virtual environment with distance estimates obtained in the corresponding physical environment. We also compare distance perception accuracy across two different exposure conditions: one in which the participant experiences the virtual space before s/he is exposed to the real space, and one in which the participant experiences the real space first. We find no significant difference in the accuracy of distance estimates obtained in the real vs. virtual environments when the virtual environment represents the same space as the occupied real environment, regardless of the order of exposure, but, consistent with previously reported findings by others, we find that distances are significantly underestimated in the virtual world, relative to the real world, when the virtual environment represents a different space from the occupied real environment. In the case of the non-co-located environment only, we also find a significant effect of previous experience in the represented space, i.e. participants who complete the experiment in the real world first exhibit less distance underestimation in the corresponding virtual environment than do participants who complete the experiment in the virtual world first. This finding is consistent with the hypothesis that participants will be more likely to act in a virtual environment in the same way as they would the real world when they are able to cognitively accept the presented virtual environment as being ‘equivalent’ to a real world space.