This research develops and implements a dyad-level procedure for testing whether the outcomes of buyer-seller negotiations correspond to settlements prescribed by Nash's theory of cooperative games. The procedure entails a multivariate statistical test in which some parameters are estimated by simulation, while others are provided directly from the assessment of negotiator utility functions by means of conjoint analysis. The procedure is applied to an experiment in which subjects participated in a realistic role-playing exercise that replicated the purchase of television advertising time. Results indicate that at the 0.01 significance level, 57.8% of the dyads can be rejected as not having achieved Nash solutions. These results, and their implications for marketing practitioners as well as future research, are discussed.