|
Activity Number:
|
109
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 12, 2002 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods*
|
| Abstract - #301903 |
|
Title:
|
Investigating the Efficacy of the Non-Interview Adjustment and Poststratification in the Accuracy and Coverage Evaluation's Housing Unit Dual System Estimation
|
|
Author(s):
|
Roger Shores*+ and Michael Beaghen
|
|
Affiliation(s):
|
U.S. Census Bureau and U.S. Census Bureau
|
|
Address:
|
4700 Silver Hill Road, 2121-2, Washington, District of Columbia, 20233, USA
|
|
Keywords:
|
simulation ; correction bias ; housing unit coverage
|
|
Abstract:
|
The Accuracy and Coverage Evaluation (ACE) produced estimates of both person and housing unit coverage for the Census 2000. This study ocuments research into the efficacy of two features of the housing unit dual system estimation, the non-interview adjustment and the poststratification. If a successful ACE person interview was not conducted at an ACE housing unit that unit did not have an ethnicity and occupancy status assignment. The non-interview adjustment spread the weights of non-interviewed housing units to those with interviews. This measure allowed for estimates of housing unit coverage by ethnicity and occupancy status and for poststratification by these variables. However, the non-interview adjustment discarded potentially useful information. This study conducts simulations to evaluate the tradeoff between gains in reduction of correlation bias versus increases in variance due to the smaller sample. It also evaluates the efficacy of the poststratification variables in combating correlation bias by re-estimating the housing unit coverage with different poststratification schemes.
|
