Repeated judgments in elicitation tasks: efficacy of the MOLE method