|
Activity Number:
|
127
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 4, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #301085 |
|
Title:
|
Relaxing the Autonomous Independence Assumption for Census Coverage Measurement Dual System Estimates
|
|
Author(s):
|
Richard A. Griffin*+
|
|
Companies:
|
U.S. Census Bureau
|
|
Address:
|
104 Arrowhead Circle, Lansdale, PA, 19446,
|
|
Keywords:
|
Dual System Estimation ; Taylor Linearization ; Sensitivity Analysis
|
|
Abstract:
|
This paper develops expressions for the expected value and variance of the dual system estimate relaxing the autonomous independence assumption. The development assumes all members of the population are exposed to possible inclusion in both the coverage measurement survey and the census. Taylor Linearization is used as in Wolter's 1986 JASA paper; however, the captures of persons in the same housing unit are no longer assumed to be independent. A sensitivity analysis is done to analyze the effect on the bias and multinomial model variance of the dual system estimate for varying amounts of capture dependence. Data from the Census 2000 Accuracy and Coverage Evaluation Survey is used to estimate potential levels of capture dependence for persons in the same housing unit.
|
