We study behavior and equilibrium selection in network games. We conduct a series
of experiments (with 580 participants) in which actions are either strategic substitutes or strategic
complements, and participants have either complete or incomplete information about the
structure of a random network. In our initial set of experiments on 5-person networks, we find a
great deal of qualitative and quantitative support for the theoretical predictions of the Galeotti,
Goyal, Jackson, Vega-Redondo, and Yariv (2010) model. The degree of equilibrium play is
striking, in particular with incomplete information. There are intriguing patterns in our data, such
as a taste for positive payoffs (but also security) when this supports the choice of one of the
potential equilibria in a complete-information setting. To shed further light on the underpinnings
of behavior and equilibrium selection in the laboratory, we study three more 5-person networks
and test robustness by conducting sessions with three 20-person networks. Overall, we see strong
evidence that choices and the equilibrium played depend on one’s degree and the connectivity of
the network, and suggestive evidence that choices also depend on the clustering in the network.