Do Groups Fall Prey to the Winner’s Curse

In an experiment, we studied how small groups tackle a company takeover game, a task where participants deciding in isolation frequently exhibit the winner’s curse. We found that groups of three members, who could exchange opinions and chat, substantially reduced the winner's curse and generally placed better bids than individuals. We report that risk attitude cannot account for the group improvement and that learning from simply observing others ' bids improved group performance only marginally. We examined the decisional processes that drove the improvement in group performance, including a detailed content analysis of group communication. When there was disagreement within a group, what prevailed was not the best but the median opinion. Hence, although groups in this task underperformed with respect to a “truth wins ” benchmark, they outperformed individuals deciding in isolation