A Method for Studying Difference in Segregation Levels Across Time and Space

Benjamin Elbers , Columbia University

An important topic in the study of segregation are comparisons over time. Theil's information index H is frequently used to study change. The H is sometimes treated as a “margin-free” measure, implying that the H is insensitive to changes in the size of racial groups or organizational units. This conclusion is not warranted, which complicates the understanding of differences in segregation levels. This paper first shows that the closely related M index has some advantages over the H. Second, by combining a method first introduced by Karmel and Maclachlan (1988) with the advantages of the M index, it is shown that a decomposition of changes in the M index is possible: one component captures changes that are due to the changing marginal distributions, and one component captures changes due to structural increases or decreases in segregation. Both components can be further decomposed to study the precise sources of changing segregation.

See paper

 Presented in Session 3. Population, Development, & the Environment; Data & Methods; Applied Demography