A method for examining learning in negotiation

The impact of feedback on interpersonal learning in negotiation was examined. An interactive computer program was developed to isolate the effect of individual judgment on performance. Subjects negotiated three times with a computerized opponent whom they were led to believe was another subject. Some subjects received a complete diagnosis of their opponents' interests following each negotiation (full feedback); others only learned about their opponents' payoffs (outcome feedback); some did not receive any information about the opponent (control). The prediction was that subjects who received a complete diagnosis would make more accurate judgments about their opponents' interests and reach more integrative agreements in subsequent negotiation situations. The results provided weak support for the model. Two indices of performance were studied: negotiators' ability to recognize compatible issues and logrolling, or the ability to make mutually beneficial tradeoffs among issues. The pattern of findings was dramatically different for the two performance measures: Whereas logrolling improved as negotiators gained experience, recognition of compatible issues worsened over trials. The degradation of performance for compatible issues was curbed for negotiators who were provided with full feedback. The feedback did not affect logrolling performance.