Operator Decision Performance Using Probabilistic Displays of Object Location

Abstract

This study compares the relative effectiveness of a conventional display of object location with the effectiveness of various probabilistic displays of object location. Several display possibilities for presenting probabilistic information to a system operator are discussed, and an experiment is performed to test the potential usefulness of selected probabilistic display types in improving operator performance. Performance evaluation metrics include the degree of accuracy achieved by the operator in estimating probabilities of events and the expected value of his decisions within a simulated search-attack mission. The effectiveness of operators using a nonprobabilistic (conventional) display also is tested by these same metrics. In general, probabilistic information processing appears to offer an improved alternative to nonprobabilistic information processing. With premission training and definitively presented probability information, probabilistic displays can be interpreted accurately by the operator and employed effectively by him within a system.