Models for Ranking Errors in Ranked Set Sampling