This paper characterizes an index that is informationally richer than the commonly used ethno-linguistic fractionalization (ELF) index. Our measure of fractionalization takes as a primitive the individuals, as opposed to ethnic groups, and uses information on the similarities among them. Compared to existing indices, our measure does not require that individuals are pre-assigned to exogenously determined categories or groups. We provide an empirical illustration of how our index can be operationalized and what difference it makes as compared to the standard ELF index. This application pertains to the pattern of fractionalization in the USA.