The paper can be found here.

Abstract
We experimentally investigate information aggregation through majority voting when some voters are biased. In such situations, majority voting can have a "dark side", i.e. result in groups making choices inferior to those made by individuals acting alone. We develop a model to predict how two types of social information shape efficiency in the presence of biased voters and we test these predictions using a novel experimental design. In line with predictions, we find that information on the popularity of policy choices is beneficial when a minority of voters is biased, but harmful when a majority is biased. In theory, information on the success of policy choices elsewhere de-biases voters and alleviates the inefficiency. In the experiment, providing social information on success is ineffective. While voters with higher cognitive abilities are more likely to be de-biased by such information, most voters do not seem to interpret such information rationally.