Group Structures

Because ethnic identities are socially and politically constructed, there is rarely one objective way of defining relevant ethnic groups in any country. Instead, different sources often identify different ethnic groups in a country, use different names for the same groups, or aggregate groups into larger “umbrella” categories. In combining different ethnic surveys within countries, the Global Diversity Index project works to identify the connections between different ways of conceptualizing ethnic identity for each country. In addition to identifying the different names given to particular groups, GDI also identifies the ways that ethnic groups have been clustered together under broader “umbrella” categories. For some countries, there may be multiple levels of aggregation, so that medium-sized clusters are also combined under even broader umbrellas.

Rather than attempting to generate an objective, fixed picture of how ethnic groups are structured in each country, the GDI extracts these structures from the set of ethnic categories included in each survey used in the project (see Ethnic Survey Sources). In attempting to correct for systematic measurement errors in these surveys, GDI relies on demographic data from national censuses (see Ethnic Census Sources). As such, the ethnic categories included in each census are also included in the group structures illustrated here. Finally, to integrate the GDI with other existing social science data on ethnic groups, categories included in All Minorities at Risk (AMAR), Composition of Religious and Ethnic Groups (CREG), Ethnic and Cultural Diversity by County (Fearon, 2003) Ethnic Power Relations (EPR), Fractionalization (Alesina et al., 2003), and Minorities at Risk (MAR) are also added to these structures.

Group structures for each country will soon be available for download from the Illinois Data Bank. For each country, two versions will be available. The first is a PDF illustration of the full structure (see image below for an example). The relevant groups are listed at each “level” of aggregation – from Level 1 (most aggregated, left-hand side) to Level N (most disaggregated, right-hand-side), with the number of levels varying from 1 to 6. At each “level,” the relevant groups are listed in bold. Beneath each group, all of its distinct names (at this level) are listed in italics, with citations added as footnotes to the relevant synonym. Subgroups of “umbrella” categories are listed to the right, at the next level of disaggregation, with appropriate citations added as footnotes to each subgroup name. In some cases, different data sources clustered groups in ways that could not be aligned, so multiple potential structures were generated.

A spreadsheet will also be available to go along with each PDF illustration, listing the group names at each level of aggregation. These spreadsheets are designed to help scholars identify ethnic structures and groups in the analysis of the text.