|
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 - #301190 |
|
Title:
|
Direct Estimates as a Diagnostic for Dual System Estimators Based on Logistic Regression
|
|
Author(s):
|
Mary H. Mulry*+ and Bruce D. Spencer and Tom Mule and Nganha Nguyen and Eric Schindler
|
|
Companies:
|
U.S. Census Bureau and Northwestern University and U.S. Census Bureau and U.S. Census Bureau and U.S. Census Bureau
|
|
Address:
|
2884 Manorwood Trail, Fort Worth, TX, 76109,
|
|
Keywords:
|
census ; Accuracy and Coverage Evaluation Survey ; 2010 Census Coverage Measurement Program
|
|
Abstract:
|
The 2010 Census Coverage Measurement Program (CCM) is preparing to use logistic regression modeling in the estimation of net census coverage error rather than post-stratification, the approach used for previous censuses. The most important objective for the CCM is to obtain separate estimates of erroneous census inclusions and census omissions. The plan for estimating census omissions is to sum estimates of net coverage error and erroneous enumerations. The net error estimates will be based on dual system estimation formed with separate logistic regression models for the correct enumeration rate and the match rate. Direct estimates at the block cluster level aid in variable selection by comparing the accuracy of estimates based on logistic regression models (or post-stratification designs) with and without a variable for groups of the clusters with different characteristics.
|
