ON THE MEASUREMENT OF POLARIZATION

Suppose that a population of individuals may be grouped according to s ome vector of characteristics into ''clusters,'' such that each cluste r is very ''similar'' in terms of the attributes of its members, but d ifferent clusters have members with very ''dissimilar'' attributes. In that case we say that the society is polarized. Our purpose is to stu dy polarization, and to provide a theory of its measurement. Our conte ntion is that polarization, as conceptualized here, is closely related to the generation of social tensions, to the possibilities of revolut ion and revolt, and to the existence of social unrest in general. We t ake special care to distinguish our theory from the theory of inequali ty measurement. We derive measures of polarization that are easily app licable to distributions of characteristics such as income and wealth.