An Age-Period-Cohort Analysis of Census Net Undercount Rates from 1940 to 2010 Using Demographic Analysis

• Activity Number: 242 - Issues in Frame Quality and Accuracy Assessments
• Type: Contributed
• Date/Time: Monday, July 29, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Government Statistics Section
• Author(s): Eric Jensen* and Lauren Medina
• Companies: U.S. Census Bureau and U.S. Census Bureau
• Keywords: Census undercount; Demographic Analysis; Age-Period-Cohort

Abstract

Demographic Analysis (DA) is one of two methods that the U.S. Census Bureau uses to estimate coverage in the decennial census. DA is a technique that combines historical vital records on births and deaths and data on international migration to develop estimates of the population that are independent of the census. Net undercount rates by age are calculated as the percentage difference between the DA estimates and the census counts for a given age or age group. In this paper, we analyze trends in census net undercount rates by age from 1940 to 2010 using the DA method. Specifically, we use Age-Period-Cohort (APC) models to analyze net undercount rates over time. These models allow us to isolate age, period, and cohort effects in the census net undercount rates or age groups that show consistent coverage issues in the census, coverage patterns that are specific to a particular census, and cohort effects that may indicate potential problems in the DA estimates. The results will inform research used to develop the final methodology for the upcoming DA and assist in evaluating the 2020 Census.