Chung Sze Kwok¹, Katelyn Becker¹, Ya Min Phyu¹, Frank H. Durgin¹; ¹Swarthmore College

Vertical extents typically appear larger than horizontal extents of the same size (horizontal-vertical illusion: HVI). The HVI may be as high as 10% for small objects viewed on a screen. But for very large objects (buildings, light poles), the magnitude of the HVI can be much larger (15-25%). Previous research using virtual environments (VEs) has revealed that medium-sized poles (0.5 -1.5 m tall) elicit a large (15%) HVI when compared to horizontal extents that are in a different depth plane, whilst only eliciting a small (10%) HVI when in the same depth plane. The present study tested whether the effect of depth separation would be observed with real objects outdoors. It additionally tested for the HVI with the objects viewed through a large overhead mirror at 45° while lying supine, using the same effective eye-height as when standing. A total of 48 participants were tested, in both upright and supine postures, using a method of adjustment: They adjusted the frontal ground distance between two balls to match the height of poles (up to 1.3 m tall). The balls were either in the same depth plane as the pole or in a nearer or farther depth plane. When in a different depth plane, matches overestimated the actual vertical by about 15%. When in the same depth plane, overestimation was not greater than 10%. Thus, these results replicate the findings previously observed using VEs, and show that the depth-separation effect can be obtained even with supine observers. An additional angle estimation task, which directly followed the HVI task, found that explicit estimates of azimuthal angular direction were more exaggerated when participants were supine than when they were standing. This latter finding is consistent with previous observations in VEs, and has now also been replicated in supine observers viewing the real world.