Index of Dissimilarity

Index of Dissimilarity The Index of Dissimilarity is a statistical measure used to quantify the degree to which two groups are unevenly distributed across subunits of a larger area. It is widely employed in studies by organizations such as the United Nations, World Bank, U.S. Census Bureau, OECD and scholars at institutions like Harvard University, University of Chicago, Columbia University, and University of California, Berkeley. The measure has been applied in analyses by researchers affiliated with Thomas Schelling, Douglas Massey, Nancy Denton, William Julius Wilson, and policy reviews by Department of Justice and National Urban League.

Definition

The Index of Dissimilarity is defined to indicate the proportion of one group's members who would need to move to different subareas to achieve an even distribution relative to another group. Classic uses compare two groups such as populations studied by U.S. Census Bureau categories, voting blocs analyzed after the Voting Rights Act of 1965, immigrants considered by International Organization for Migration, and socio-demographic studies at Brookings Institution. The index ranges from 0 (complete evenness) to 1 (complete segregation), and has been interpreted in public debates involving actors like Martin Luther King Jr., Malcolm X, Bayard Rustin, and organizations including NAACP, ACLU, and Urban Institute.

Calculation and Properties

Computation of the index uses counts at the subunit level such as neighborhoods, census tracts, wards, or school districts used by authorities like U.S. Census Bureau, ONS (Office for National Statistics), Statistics Canada, Eurostat, and research by Rand Corporation. For each subunit i, let A_i and B_i denote counts of two comparison groups; the index is derived from the sum across i of |A_i/N_A − B_i/N_B| where N_A and N_B are total group sizes. This arithmetic mirrors approaches in earlier quantitative work associated with Alfred Lotka, Harold Hotelling, Ronald Fisher, and is analogous to metrics used in transportation studies by John von Neumann-adjacent research. Properties include scale invariance under proportional changes, sensitivity to spatial aggregation similar to the Modifiable Areal Unit Problem discussed by researchers at MIT and University College London, and interpretability that made it useful in analyses by Harvard Kennedy School and policy reports from Congressional Research Service.

Applications

Scholars and practitioners have applied the index in urban studies at University of Pennsylvania, public health research at Johns Hopkins University, education policy analyses at Teachers College, Columbia University, and criminal justice studies cited by Brennan Center for Justice. Case studies include segregation assessments in Chicago, New York City, Los Angeles, London, Paris, Johannesburg, and São Paulo. It has informed litigation under the Voting Rights Act of 1965, zoning debates involving Boston, transit planning in San Francisco, school districting in Atlanta, and comparative international reports by UN-Habitat and World Health Organization.

Limitations and Criticisms

Criticism of the index has come from scholars at Princeton University, Yale University, Stanford University, and University of Michigan who highlight issues like insensitivity to spatial arrangement (it ignores adjacency), dependence on the number and boundaries of subunits—echoing concerns raised in studies by David Harvey and debates at the American Sociological Association—and inability to account for multi‑group contexts without pairwise decomposition, noted in critiques by Iceland, Weinberg, and Steinmetz-style analyses used by think tanks such as Migration Policy Institute. Legal scholars at Harvard Law School and Georgetown University Law Center have debated its applicability in court under the Fourteenth Amendment and in remedies shaped by cases like Brown v. Board of Education and Shelby County v. Holder.

Variants and Related Measures

Related indices include the Gini coefficient used in income inequality studies at World Bank and OECD, the Theil index developed in information theory contexts at Bell Labs and discussed by Shannon, spatial autocorrelation measures like Moran's I applied by geographers at University of Chicago, and clustering metrics such as the Isolation index and Exposure index employed in demography courses at London School of Economics and University of California, Los Angeles. Other spatially explicit approaches include kernel density estimation used in research at University of Washington and local indicators of spatial association popularized by scholars at National Center for Geographic Information and Analysis.

Category:Segregation metrics