We present a parametric learning model of players' dynamic and possibly out-of-equilibrium beliefs about other players' preferences that also incorporates random utility (noise). We estimate the model using the data from the four-country ultimatum game experiments of Roth et al. (1991). We find...