Accuracy of Pointing to Invisible Landmarks in a Familiar Environment

Humans develop a mental representation of their environment that enables them to solve their day-to-day navigation tasks. The nature and quality of this representation can be probed experimentally by studying spatial tasks like distance and bearing estimation. We investigated pointing accuracy in an outdoor environment as part of a larger project in which we compare spatial behaviour in real environments with that in virtual environments (virtual reality).

We carried out experiments in the centre of Tübingen (ca 600 m × 400 m). Standing near one of eleven well-known landmarks, subjects had to turn a pointer in the estimated direction of each of the other ten invisible (occluded) landmarks (all ten subjects had lived in Tübingen for at least two years). The estimated directions were read by the experimenter from a hidden compass and later compared with the actual directions. This procedure was repeated from all eleven locations along the subjects' route through the town.

The mean absolute angular error of pointing was 10.8°, while the individual subject's means ranged from 9.1° to 15.0°. We found highly significant differences in pointing accuracy for the different locations; mean absolute error ranged from 6.3° to 15.8° (n=10). Systematic asymmetries in pointing from one landmark to another and back were not found.

The pointing accuracy found in this study is clearly better than that found in most other studies, which might relate to the familiarity and the size of the environment. The apparent absence of an asymmetry in the pointing directions between two locations is consistent with the notion of a metric mental map.