How to Collect Private Signals in Information Cascade: An Empirical Study

In the information cascade experiment, several subjects sequentially answer a two-choice question, after referring to previous subjects’ choices. Information cascade is defined as a tendency to follow the majority choice, even if, one’s private signal suggests the minority choice. When information cascade occurs, the private signal is lost, and the collective intelligence mechanism does not work. If the majority’s choice is wrong at the onset of the information cascade, it continues to be wrong forever. How can we find the correct choice even when the majority choice is wrong? In this study, we investigate a Bayesian Inference method, which collects private signals in the information cascade, based on the choice behavior of the subjects. Using the empirical data of an experiment, we estimate the probabilistic rule of the choice behavior. We demonstrate that the Bayesian algorithm works and one can know the correct choice even if the majority’s choice is wrong.