242 – Issues in Frame Quality and Accuracy Assessments
Imputation Models Using Computer Matching Results
Glenn Reisch
U.S. Census Bureau
The major goal of the 2010 Census Coverage Measurement (CCM) program was to measure coverage for housing units and people in the 2010 Census using dual-system estimation. As part of this process, person interviews were conducted at a sample of housing units collected independently of Census operations. After computer matching of this data to Census results, extensive clerical review and field followup were performed on nonmatches and cases with unresolved statuses. The 2020 Post-Enumeration Survey (PES) has the same goals and approaches as the 2010 CCM, which includes clerical matching and field followup. However, in this research, we investigated using models instead of clerical review or followup operations. Not including clerical review or followup operations would lead to larger numbers of nonmatches and unresolved statuses. We used models and information from the 2010 CCM to impute the match status and unresolved statuses. We then used the final 2010 CCM results to evaluate coverage estimates based on computer matching output.