Abstract
Background
Ethnic density is widely used in epidemiology and health geography as a contextual exposure, usually defined as the proportion of residents who share an individual’s ethnic group. Yet it is rarely examined as a measurement problem in its own right, and equivalent percentage values may represent very different neighbourhood contexts across groups and places.
Methods
Using the new harmonised Unified UK Census Data release, I analysed 239,023 small areas across England, Wales, Scotland and Northern Ireland. I treated ethnic density as a contextual construct to be decomposed, deriving ethnic density measures from the harmonised ethnicity table and theory-led explanatory domains from harmonised tables on migration, language, religion, household and age structure, housing, and socioeconomic position. I examined ecological correlates, estimated sequential decompositions, and assessed regional embeddedness and cross-regional variation in selected correlational profiles.
Results
Ethnic density did not behave as a single contextual scalar. White density aligned strongly with UK-born composition, Asian density with Middle East/Asia-born and Muslim composition, Pakistani density with Muslim share and large households. Migration explained most variance in White (R2=0.890), Asian (R2=0.850), and Black (R2=0.729) density, whereas religion contributed the largest additional gain for Pakistani (ΔR2=0.321) and Indian (ΔR2=0.251). Regional embeddedness also varied markedly across groups, indicating that some ethnic density measures reflect broader settlement systems as well as local neighbourhood composition.
Conclusions
Ethnic density is better understood as a group-specific and spatially embedded ecological bundle than as a singular contextual exposure. Equivalent percentage values are not necessarily comparable across ethnic groups, with implications for interpretation, adjustment strategies, and the risk of over-adjustment in ethnic density models.
Key messages
• Ethnic density measures different neighbourhood contexts across ethnic groups.
• Ethnic density models need clearer estimands and more cautious adjustment.