This paper reports the results of a two-part data analysis of learning in a repeated costly decision experiment. In the first part we test payoff dominance under the hypothesis of expected payoff maximization. We utilize a dynamic probability distribution over decisions for each player,...