Methods for Conducting An Administrative Records Census Experiment in 2000

• Activity Number: 276
• Type: Topic Contributed
• Date/Time: Wednesday, August 14, 2002 : 8:30 AM to 10:20 PM
• Sponsor: Section on Government Statistics*
• Author(s): Charlene Leggieri*+ and Arona Pistiner
• Affiliation(s): U.S. Census Bureau and U.S. Census Bureau
• Keywords: administrative records ; census ; experiment

Abstract

The Census Bureau's Administrative Records Experiment required the development of previously untested methods for merging national sets of administrative data into a statistical system simulating a census type accounting of the population (block level counts with age, gender, race/ethnicity). The purpose of the experiment was to test the feasibility of using administrative records as a primary data collection method for conducting the census. Seven national administrative files from the Internal Revenue Service, Social Security Administration, Health Care Finance Administration, Selective Service, HUD, and Indian Health Service were edited and merged into a Statistical Administrative Records System (StARS). Two methods of simulating a census were tested. One method geocoded the address information from the unduplicated lists to census blocks and tallied the results. The second method geocoded records to the block but also attempted to reconcile addresses within a frame of residential addresses (the Master Address File). The paper describes the methods used for constructing StARS and operational aspects of the experiment, along with preliminary results of the national system.