|
Abstract:
|
Preparing for the 2020 Census, the U.S. Census Bureau is researching the use of Undeliverable-as-Addressed mail (UAA) from the United States Postal Service in the prediction of vacant and non-existent housing units in the Nonresponse Followup universe. This is part of an effort to reduce the cost of the 2020 Census while maintaining data quality. For the majority of addresses, the Census Bureau sends multiple mail pieces to encourage response, such as an advance letter and internet invitation postcard. When these mailings are returned as UAA, the specific reason assigned by the postal carrier provides useful information about the status of the address. In this paper, we look at the role of UAA by considering patterns and consistencies in the assignment of UAA reasons. We discuss possible techniques for accounting for inconsistencies in UAA assignments across mailings such as including Cohen's kappa coefficient for agreement as a covariate for modeling housing unit status. This paper also considers changes in the national UAA assignment rate over time and its implications for our rate of vacancy assignment.
|
