Thirty-six two-man teams participated in a negotiation experiment which used 2 x 2 game matrices. An attempt was made to shape the negotiating behavior by manipulating the prior experiences of the teams. There were three experimental conditions. Each was designed to foster one particular orientation toward the negotiations. The intended orientations were (a) competitive, (b) cooperative, and (c) limited trust. No systematic differences were found, however, between the teams in the three conditions. The implication of this results, with particular reference to the position forwarded by Blake and Mouton, was briefly discussed. In a second session using the same teams, differently paired and in different experimental conditions, a study was made of the differences in the perception of the negotiations by participants and nonparticipants (observers). Several notable differences were found which highlighted the arbitrariness of experimental procedures which depend on either self-report or observational rating, but not both.