Overload and Confirmation Bias

We show that stronger information overload strengthens confirmation bias. In an experiment, we keep constant the information received by subjects but vary the difficulty of information processing. In the treatment condition, the preceding signal disappears as the next signal appears. In the control condition, the preceding signal remains visible. It is thus more difficult to process information in the treatment condition, i.e., there is stronger information overload in the treatment than in the control condition, which is supported by the fact that subjects make more belief updating mistakes in the treatment condition. We find stronger confirmation bias among subjects in the treatment condition, in particular when they receive belief-challenging information. Our results provide empirical support for recent theories that emphasize the role of limited ability and complexity, and caution against information campaigns that ignore the role of cognitive load and attention.