When decisions are made sequentially in a group, Bayesian playersaiming to maximize individual payoffs often need to ignore their private information and imitate choices of earlier decision makers. This results in information cascades. Once formed, information cascades can be harmful to information aggregation and average payoffs in the group. Therefore, there is a potential dilemma facing players in such situations. They can reveal private information and risk lower immediate payoffs but benefit the group in the long run, or join an information cascade. It has been shown that reciprocity (i.e., paying a cost at one time while being compensated at another) can lead to optimal solutions in such dilemmas. We hypothesized that it would be so for the present one, as well. This hypothesis was tested in an experiment in which we manipulated two factors: the stability of players' positions in a sequence that determined whether reciprocity was possible, and the decision payoff structure that constrained to what extent revealing private information was beneficial to the group. The results show that both factors influenced the probabilities of private information revealing and the group payoffs in ways consistent with our predictions.