Activity Number:
|
278
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 14, 2002 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Survey Research Methods*
|
Abstract - #301907 |
Title:
|
Overview of the U.S. Census 2000 Direct Variance Estimation
|
Author(s):
|
Lisa Fairchild*+ and Philip Gbur
|
Affiliation(s):
|
U.S. Census Bureau and U.S. Census Bureau
|
Address:
|
14205 MacFarlane Green Ct. #4135, Upper Marlboro, Maryland, 20772, USA
|
Keywords:
|
Systematic Sample ; Successive Differences
|
Abstract:
|
For the U.S. Census 2000, a systematic sample of addresses across the country received a long form, or sample, questionnaire which collected detailed socioeconomic and demographic characteristics. The selected addresses and the people at those addresses were weighted to represent the entire population and housing stock of the nation. Direct variances were estimated for a subset of long form estimates which were then generalized to produce variance estimation parameters for use with all long form estimates. The successive difference replication variance estimation methodology was used for the direct estimates. This paper will present a brief description of the sample design and weighting methodology for the long form questionnaire and an overview of the direct variance estimation methodology.
|