DISTRIBUTED DETECTION IN TEAMS WITH PARTIAL INFORMATION - A NORMATIVE-DESCRIPTIVE MODEL

This paper considers a hierarchical team faced with a binary detection problem, wherein decision makers (DM's) have access to different subs ets of noise-corrupted information about the true state of the environ ment. A normative model is developed that aggregates individual expert ise of DM's at different levels of hierarchy. The resulting team exper tise is characterized iii the form of a Team Receiver Operating Charac teristic (ROC) curve, thereby replacing the team by an equivalent sing le decision-making node. The normative model is tested against human t eams in a laboratory experiment. The team objective is to minimize the cost of errors in the final decision at the primary DM, where the cos t structure and the information structure are treated as independent v ariables. Discrepancies between normative predictions and experimental results are attributed to inherent limitations and cognitive biases o f humans. These human characteristics are quantified and the normative model is augmented with psychologically interpretable (descriptive) f actors. The resulting normative-descriptive model yields accurate pred ictions of both the performance and strategy variables of human teams.