Citation: Casella, A., Kartik, N., Sanchez, L., & Turban, S. (2018). Communication in context: Interpreting promises in an experiment on competition and trust. Proceedings of the National Academy of Sciences. http://doi.org/10.1073/pnas.1714171115
Abstract: Are there differences in learning when information is ambiguous relative to when it is not? This paper explores how the introduction of ambiguity in a social learning game affects the strategies chosen by players. We test the multiple priors model in a laboratory experiment of informational cascades. Our findings suggest a high level of probabilistic sophistication in part of our subjects and provide limited support for the multiple priors hypothesis. Although a substantial number of subjects exhibit the traditional ambiguity averse preferences in an independent elicitation task, we conclude that ambiguity has little to no effect in social learning even when we restrict attention to ambiguity averse subjects.
PDF Download: Communication in Context (PNAS Final)