In a norm-referenced assessment a candidate’s success or failure depends on the performance of the other examinees. Theoretically, identical scores attained by candidates in sequential high- and low-achieving year groups might fail one year but pass the next. In contrast, in a criterion referenced exam, a candidate’s achievement is assessed independently of the performance of others by using clearly stated standards related to predetermined learning outcomes. All you need to do is to define with absolute clarity that fine and unwavering line across each skill, every element of knowledge and nuance of attitude that once rated and ordered, differentiates ‘a doctor’ from ‘not a doctor’. Current dogma is that if you cannot make criterion referencing work you are not trying hard enough.