Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
The payoff of actions is estimated and the resulting empirical payoff is controlled for in regression analyses to formulate a test of rational expectations in information cascade experiments. We show that the empirical payoff of actions is a function of estimates of choice probabilities and estimates of the information parameters of the game. We introduce an alternative empirical payoff of actions with true values of the information parameters. Our improved measure of the success of social learning confirms that rational expectations are violated, but deviations from rational expectations are statistically significantly smaller than in Weizsacher (2010).