an electorate. We propose a novel model of these underlying social interactions, and derive
maximum likelihood estimators for the ground truth in these models, given the social network
and votes. We also evaluate these new estimators, as well as existing ones, on a class of
simulated social networks.