The Effect of Biased Communications On Both Trusting and Suspicious Voters

In recent studies of political decision-making, apparently anomalous behavior has been observed on the part of voters, in which negative information about a candidate strengthens, rather than weakens, a prior positive opinion about the candidate. This behavior appears to run counter to rational models of decision making, and it is sometimes interpreted as evidence of non-rational "motivated reasoning". We consider scenarios in which this effect arises in a model of rational decision making which includes the possibility of deceptive information. In particular, we will consider a model in which there are two classes of voters, which we will call trusting voters and suspicious voters, and two types of information sources, which we will call unbiased sources and biased sources. In our model, new data about a candidate can be efficiently incorporated by a trusting voter, and anomalous updates are impossible; however, anomalous updates can be made by suspicious voters, if the information source mistakenly plans for an audience of trusting voters, and if the partisan goals of the information source are known by the suspicious voter to be "opposite" to his own. Our model is based on a formalism introduced by the artificial intelligence community called "multi-agent influence diagrams", which generalize Bayesian networks to settings involving multiple agents with distinct goals.