Robinson's Coefficient of Agreement—A Critique

Abstract

Despite a firm conviction that statistical analyses have a definite place in archaeological interpretations, I want to register a protest against the Coefficient of Agreement proposed by W. S. Robinson (American Antiquity, Vol. 16, No. 4, pp. 293-301, 1951). Robinson's coefficient is offered as a measure of the degree of similarity between two collections of sherd material. It is arrived at by adding the differences between the percentages of each type of sherd in the two collections and subtracting this sum from 200.