This paper develops a measure of segregation based on two premises: (1) a measure of segregation should disaggregate to the level of individuals, and (2) an individual is more segregated the more segregated are the agents with whom she interacts. Developing three desirable axioms that any segregation measure should satisfy, we prove that one and only one segregation index satisfies our three axioms, and the two aims mentioned above; which we coin the Spectral Segregation Index. We apply the index to two well-studied social phenomena: residential and school segregation. We calculate the extent of residential segregation across major US cities using data from the 2000 US Census. The correlation between the Spectral index and the commonly-used dissimilarity index is .42. Using detailed data on friendship networks, available in the National Longitudinal Study of Adolescent Health, we calculate the prevalence of within-school racial segregation. The results suggests that the percent of minority students within a school, commonly used as a substitute for a measure of in-school segregation, is a poor proxy for social interactions.
