Abstract
People's estimates of distance in sparse, dark visual environments can be very poor. Gogel (1972) described a tendency to judge distances to single targets presented in the dark as if the targets were located near a single specific distance regardless of actual distance, and Foley (1985) found substantial interactions of distance estimates when even a single extra target was added to the visual scene. Foley described a model based on misestimated binocular parallax and veridical relative disparity. When attempting to replicate Foley's results, we found that (1) observers can use relative disparity in rather different ways depending on the task requirements (manual or verbal judgements) and that (2) observers performance is consistent with underestimation or overestimation of large disparities, whereas their behaviour is consistent with veridicality for small disparities (2 degrees or less). We present here an extension of Foleys model that can better describe observers use of disparity for distance estimates to targets within arms reach.