The effect of landmarks on route-learning in a computer-simulated environment

Abstract

In a route-learning task in a computer-simulated environment, subjects repeatedly negotiated a series of rooms each containing two exit doors, and were required to learn which of the doors lead to the next room. In the landmark condition each room contained distinctive objects, while in the non-landmark condition all rooms were identical. The results of experiment 1 revealed that the route-learning performance of both groups was comparable. It was hypothesized that the landmark group relied primarily on paired associate learning in which the landmark objects were associated with the correct door while the non-landmark group learnt a list of correct left/right decisions. In an attempt to suppress the latter strategy, in experiment 2 two groups of subjects were asked to perform the same tasks as the groups in experiment 1, but were required to do a backcounting task while learning the route. This manipulation resulted in superior performance on the part of the landmark group. It is argued that landmarks are only one of many successful strategies that people have at their disposal for learning routes, and that in order to show that landmarks can facilitate route-learning it is necessary to suppress other strategies.