#05-10
Learning in Networks: An Experimental Study
Syngjoo Choi, Douglas Gale and
Shachar Kariv
Abstract: This paper reports an experimental test of the model of social learning in networks described in Gale and Kariv (2003). We focus on three-person networks and investigate: the complete network, in which each subject can observe the other two subjects, and two incomplete networks, the circle, in which each agent observes one other agent, and the star, in which the subject in the center of the star is connected to the two subjects at the periphery. Learning in networks is challenging because of the lack of common knowledge about the history of play. Subjects must draw inferences about the actions other subjects have observed as well as about their private signals. Nonetheless, the rationality and efficiency of subjects’ behavior is strikingly consistent with the theory. To account for errors in subjects’ behavior, we adapt the model of Quantal Response Equilibrium (QRE) of McKelvey and Palfrey (1995, 1998) and find that its restrictions are also confirmed.
Keywords:
JEL classifications : D82, D83, C92
