Testing the Value of Route Directions Through Navigational Performance

Route directions to reach a target point on a campus were collected from undergraduates. A "good" description and a "poor" one were selected, based on ratings provided by judges in terms of their value for navigational assistance. People unfamiliar with the campus were then required to navigate to the target after studying one of these descriptions. In addition, a "skeletal" description, which contained the essentials needed for navigating, was constructed and used in the experiment. During navigation, we measured the frequencies of stops and of directional errors (whether these errors were self-corrected or corrected by the experimenter). Overall, the good and the skeletal descriptions resulted in better performance than the poor one. Their value as navigational aids was confirmed by measuring the navigation times. Analyzing the structure and content of the descriptions confirmed that the effectiveness of route directions depends on their ability to connect actions to landmarks, that is, to closely link the prescriptive and the descriptive parts of this specific type of spatial discourse.