Traditionally, the Concepts of Over- and Underachievement are defined, operationally, assuming that perfect relationship between the measures of criterion and predictor variables is approachable. This assumption is not empirically tenable. Since the measures of predictor and criterion variables are, by way of necessity, obtained on questionably reliable tools, the assumption of perfect correlation would remain an unrealizable phenomenon. The present study reveals the inadequacy of this untenable theoretical assumption underlying the traditionally popular operations and also proposes a scientifically sound method of identifying over- and underachieving pupils.