The Impact of Using Hand-Held Computers To Reduce Count Imputation in the 2004 Census Test

• Activity Number: 529
• Type: Topic Contributed
• Date/Time: Thursday, August 10, 2006 : 10:30 AM to 12:20 PM
• Sponsor: Section on Government Statistics
• Author(s): Jamie Burnham*+ and Rosemary Byrne
• Companies: U.S. Census Bureau and U.S. Census Bureau
• Keywords: monitoring nonresponse ; nonresponse followup ; computer assisted interviewing ; mobile computing environments

Abstract

The 2004 Census Test used hand-held computers (instead of paper questionnaires as in Census 2000) to collect data for personal visit interviews for the Nonresponse Followup (NRFU) Operation. During the NRFU Operation, enumerators followed up households that did not return their 2004 Census Test questionnaires. In 2000, hot deck imputation was used to impute housing unit status, household occupancy status, and/or household size when not reported or when there were data inconsistencies. The 2004 plan to reduce the need for count imputation consisted of using automation to more closely monitor addresses missing this information, using built-in edits in the questionnaire, and collecting status and population count for refusal situations. This paper presents the results of these changes on 2004 count imputation, compared to the Census 2000 rates in the corresponding NRFU areas.