Familiarity Effects and Questioning Biases in Human Belief Revision