?

What is the implicit assumption the rater makes when they make the "similarity error" in performance evaluations?