Dr. Lerch describes how self identification of race was reported in a community recognized as an Indian tribe by the state of North Carolina but not by the Federal government. Causes of coverage error in this rural community were complex. Flexible residential arrangements spread families into physically distinct structures which, situationally, could be classified as separate housing units or merely as extra bedrooms disconnected from the main house. How the boxes into which mail was delivered line up with the locations of associated housing units in the community's rural postal address system was neither consistent nor immediately apparent without tapping into community knowledge. Problems interpreting symbols on the census TIGER map caused minor geocoding errors for both census and alternative enumerators. Some differences between the Alternative Enumeration and the Census Enumeration reflect residential mobility within the sample area and within this densely inter-related kin community.
Citation: Lerch (August 1992) Ethnographic Evaluation of the 1990 Decennial Census Report # 20. Final report for Joint Statistical Agreement 89-25 with the University of North Carolina at Wilmington; September 17, 1992, PREM # 181. Dataset: LER-A NOTE: The North Carolina ethnographic site was one of four where the 1990 Census, an Alternative Enumeration, and the 1990 Post Enumeration Survey (PES-at zero weight) were conducted and compared through a triple match. The author's comments on the triple match of records originating in her Alternative Enumeration to records keyed from census forms and to Post Enumeration Survey records is entitled "A Comparison of the Alternative Enumeration, the Post Enumeration Survey, and the 1990 Census of a Rural Minority Community" and was submitted July 31, 1992. For summary results see Hamid and Brownrigg available on this server.
