Two friends of mine here at Duke published an interesting paper....
Public information and electoral bias
Curtis Taylor & Huseyin Yildirim
Games and Economic Behavior, forthcoming
We present a theory of voting that predicts that elections are more likely to be close, and voter turnout is more likely to be high when citizens possess better public information about the composition of the electorate. These findings suggest that providing more information to potential voters about aggregate political preferences (e.g., through pre-election polls or expert forecasts) may undermine the democratic process. Our analysis reveals that if the distribution of political preferences is common knowledge, then the unique type-symmetric equilibrium leads to a stark neutrality result in which each alternative is equally likely to win the election. By contrast, when citizens are ignorant about the preference distribution, the majority is more likely to win the election and expected voter turnout is lower. Welfare is, therefore, unambiguously higher when citizens possess less information about the preference distribution.
(Nod to Kevin L)