Martin Kocher, Matthias Suffer, Florian Wakolbinger
We study the impact of social learning on the depth of reasoning in an experimental beauty-contest game. Naive advice and observation of others' decisions as two forms of social learning trigger faster convergence to the equilibrium. We find that subjects who receive advice outperform uninformed subjects permanently, whereas subjects who observe others' past behavior before making their decision only have a temporary advantage over uninformed subjects. A series of control-treatments and simulations indicate that the latter result is due to subjects failing to make the most out of observing others.