The curse of uninformed voting: An experimental study

We study majority voting over two alternatives in small groups. Individuals have identical preferences but are uncertain about which alternative can better achieve their common interest. Before voting, each individual can get informed, to wit, buy a valuable but imperfect signal about the better alternative. Voting is either voluntary or compulsory. In the compulsory mode, each individual can vote for either of the two alternatives, while in the voluntary mode they can also abstain. An uninformed random vote generates negative externalities, as it may override informative group decisions in pivotal events. In our experiment, participants in groups of three or seven get informed more often with compulsory than voluntary voting, and in this way partly counteract the curse of uninformed voting when they cannot avoid it by abstaining. Surprisingly, uninformed voting is a common phenomenon even in the voluntary mode! A consequence of substantial uninformed voting is poor group efficiency in all treatments, indicating the need to reconsider current practices of jury and committee voting.