Contributors to Census omission rates include both individual and environmental factors which are associated with high rates of poverty. One related factor, racial/ethnic group affiliation, has been linked to behavioral causes of undercount. Moreover, geographic concentration of racial/ethnic groups, in either urban or rural settings, has been linked to the quality of census coverage overall. (Brownrigg and de la Puente, 1992; Fein, 1990). A number of researchers have suggested that group concentration may indeed exert independent effect: upon census omission and census coverage. Similarly, it is widely held that poverty status contributes to behavioral causes of undercount, and that geographic concentrations of poverty are linked to problems in census coverage. Analysis of the separate contributions of poverty status and race/ethnicity to the quality of coverage and the extent of undercount is complicated by the observation that geographic concentration of racial/ethnic groups often overlap with geographic concentrations of poverty. Where racial/ethnic characteristics overlap with poverty factors, overall omission rates may tend to increase.

This research examines variation in racial/ethnic heterogeneity of neighborhoods, and census omission. More specifically, the research examines variation in Census omissions for small areas while controlling for percentage composition of persons who self-identified as black. The research utilizes data from the 1990 Decennial Census, and data from the 1990 Ethnographic Census. The Ethnographic Census is designed to overrepresent pockets of poverty and targets underrepresented groups. It therefore serves as an independent sample for both poverty and race/ethnic concentration.