This paper is concerned with the measurement of agreement between two observers classifying items into nominal categories, with one of the observers being viewed as the "standard". An asymmetric version of Cohen's Kappa is proposed as an appropriate measure. Properties of this measure are outlined, and a numerical example is given.