An Analysis of Person Duplication in Census 2000

• Activity Number: 100
• Type: Topic Contributed
• Date/Time: Monday, August 9, 2004 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Robert E. Fay*+
• Companies: U.S. Census Bureau
• Keywords: census duplication ; erroneous enumeration ; computer-matching

Abstract

Computer-matching on name and date of birth can be used to identify duplicate enumerations of persons in Census 2000. Computer-matching faces two important limitations. First, because name and date of birth are not always unique, computer-matching links some enumerations together that do not represent duplicated persons. Second, as a consequence of inaccurate or missing names or days of birth, computer-matching fails to identify some duplicates. Two previous papers presented probabilistic models to address the problem posed by coincidental sharing of date of birth by persons with the same name. The paper describes the application of the methods to Census 2000. Although estimates of duplication in Census 2000 are already available from a previous study, the earlier results are based on a sample whose size limits detailed analysis. Instead, the paper reports results from the full census. In addition to applying the probabilistic models described previously, the paper investigates the degree of improvement in matching resulting from a series of edits of the reported names developed by other U.S. Census Bureau researchers.