Models of learning have provided a unified explanation of a wide variety of experimentally observed behavior. One of the subtle predictions of learning models is that the relative speed at which different players in a game adjust their behavior when they are inexperienced may have a critical influence on the long term behavior in the game. This is how learning models account for the characteristic behavior observed in ultimatum games, in which outcomes are closer to equal splits than to the simple predictions of perfect equilibrium. However ultimatum games have proved a difficult test for theories of learning, because no experiment to date has observed learning on the part of the responders, who must accept or reject an ultimatum division of the available wealth. This is not completely surprising, since the learning models predict that, because of the different incentives the game gives to proposers and responders, responders will learn very much more slowly than proposers. The present paper reports an experiment designed to test whether relative speed of learning has the predicted effect, by manipulating the amount of experience accumulated by proposers and responders. The experiment allows the predicted learning by responders to be observed, for the first time. The speed at which responders learn responds to the speed at which proposers learn, in the predicted manner.
Cooper, David, Nick Feltovich, Alvin Roth, and Rami Zwick (2003), "Relative versus absolute speed of adjustment in strategic environments: responder behavior in ultimatum games", Experimental Economics 6 (2), pp. 181-207. DOI: 10.1023/A:1025309121659