Contingent social utility in the prisoners’ dilemma

We examined a central assumption of recent theories: that social utility is contingent on impressions of other people. We manipulated participants’ impression of the other player in a prisoners’ dilemma. We then measured participants’ own preferences in the PD, their estimates of the other player’s preferences in the PD, their prediction of the other player’s move, and their own move. We hypothesized that the participants’ move would maximize their stated preferences, given their prediction of the other player’s move, and that participants’ preferences would be contingent on their impression of the other player. Results supported both hypotheses and revealed that participants’ preferences were contingent more on their estimate of the other player’s preferences than on their prediction of the other player’s move. © 2001 Elsevier Science B.V. All rights reserved. JEL classification: C72 (non-cooperative games); C91 (laboratory experiment, individual behavior)