Paying a price for information: Non-optimal decision making and preference for information over expected value outcomes

Abstract

The information purchase bias describes a tendency to opt for information that reduces uncertainty, even when the cost of information is associated with a substantial net reduction in expected value. The current experiment quantified the amount of information that could be purchased in terms of increased certainty of finding a target (a downed aircraft) by purchasing a more expensive, but more powerful sensor. In a “good deal” block, the expected financial gain of finding the target with the upgraded purchase clearly exceeded the cost of the purchase. In a neutral deal block the two terms were equal and in the bad deal block the purchase cost far exceeded the expected gain. The neutral deal was purchased 65% of the time (15% bias); and the bad deal purchased 50% of the time, where the optimal strategy would have been to never purchase. Implications for automation and training are discussed.