Files
Abstract
The Proximity Compatibility Principle (PCP) states that a display format is wellsuited to a given task if the information sources in the display are related to the same degree as information sources in the task. While experiments have shown that PCP can provide useful display design guidelines for many types of tasks, diagnosis tasks have not seemed to conform to PCP's predictions. The current experiment compared performance and user preference with integral, configural, and separable displays in three diagnosis tasks based on a medical diagnosis system. Tasks differed in variable count and differential utility. As predicted, the integral display was superior in tasks without differential utility. Differential utility diminished but did not eliminate the integral display's superiority. Users preferred the integral display and disfavored the separable display in all task conditions, especially in tasks without differential utility. The results indicate that PCP is a useful theory for diagnosis tasks, but different diagnosis tasks can differ widely in their task demands.