Transfer of spatial information from a virtual to a real environment

Participants explored either a real multistory building or a to-scale computer simulation of the same building. Following exploration, we assessed their knowledge of the spatial properties of the environment. Participants in a control group were asked to complete the same assessment tasks without the opportunity to explore either the real building or the computer simulation. Spatial knowledge was assessed using four measures: pointing to objects not directly visible from the test site, Euclidean distance estimates, route distance estimates, and drawings of the building. Transfer of learning from the simulation to the real environment was evident. Pointing to objects not visible from the test site proved to be the most sensitive measure; map-drawing accuracy was very similar to pointing accuracy. However, some other measures did not distinguish the estimates of control participants from the group that explored the real building.