288 – Census Coverage Measurement
Once and Only Once: Searching Near and Far for Person Duplications in the 2010 Census
Aaron Cantu
U.S. Census Bureau
Susanne L. Johnson
U.S. Census Bureau
Census respondents often have their own ideas regarding where people should be counted, which can lead to people being enumerated in more than one place (duplicated). Problems in establishing a person’s Census Day residence led to the 2000 coverage measurement program underestimating erroneous enumerations (many of which were duplicates). This paper will discuss methodology used by the 2010 Census Coverage Measurement (CCM) Program to improve measurement of Census Day residence and improve techniques to detect and resolve duplicate enumerations. CCM collected address information for other places people could have been counted on Census Day. We conducted computer and clerical searches for census duplicates and matches between the CCM and census near these additional addresses (on top of searching near the sample address). We also conducted a nationwide computer search that was less conservative than previous computer searches for census duplicates because CCM had the opportunity to clerically review these computer links and conduct field followup, when necessary, to resolve whether the links were truly duplicates and establish where the person should have been counted.