AbstractA 60‐trial iterated PDG was played by 24 male undergraduates. Half were Abstract Information Processors, and half were Concrete Information Processors. In addition, half of the subject pairs played the game face‐to‐face, while in the other pairs the game was played in separate cubicles. Rather than allowing for free play, the subjects were given the illusion of playing against one another when in reality they all played against a simulation program. It was found that concrete subjects cooperated most in the presence of another player, and competed most when not facing him. This was apparently due to the tendency for the concrete players to be “caught” by the cooperative pull of the face‐to‐face condition. The abstract subjects tended to use information‐seeking strategies in both conditions. When the information was irrelevant to the game (e.g., presence of a false “partner”) the abstract players did not maximize to the same degree as when the feedback