We present an experiment to compare the two competing fairness theories by Bolton and Ockenfels and by Fehr and Schmidt. For most experiments that these theories have previously been applied to, they make similar predictions so that it is difficult to compare their predictive accuracy. We designed a very simple experiment that induces opposing predictions made by both theories. If the maximization of total payoff is in line with the decision predicted by Fehr and Schmidt, almost all subjects decide accordingly, whereas if it is in line with the prediction by Bolton and Ockenfels, decisions are dispersed. Hence the performance of the theory by Fehr and Schmidt is much better than that of the theory by Bolton and Ockenfels, although both theories ignore the importance that subjects assign to efficiency