|
Activity Number:
|
248
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #302887 |
|
Title:
|
Small Area Estimation for Population Counts in the German Census 2011
|
|
Author(s):
|
Ralf T. Münnich*+ and Jan P. Burgard and Martin Vogt
|
|
Companies:
|
University of Trier and University of Trier and University of Trier
|
|
Address:
|
Economic and Social Statistics Department, Trier, International, 54286, Germany
|
|
Keywords:
|
small area estimation ; spatial modeling ; mixed logistic models ; register-based Census ; undercounts ; overcounts
|
|
Abstract:
|
In modern survey applications, National Statistical Institutes have always the pressure to reduce costs. This task plays an important role in the next European Census round 2010/11, where some countries try to employ a register-based Census which may help to reduce costs by far. Small area estimation methods are expected to allow for high-level results for small areas and domains under the given budget constraints. The present article focuses on the estimation of higher educated people as example for a variable which is not overserved in the register. Based on the classical basic unit and area-level models, binomial mixed-models will be elaborated. Finally, spatial small area models will be assessed. The entire study will be accompanied by a Monte-Carlo study which will foster comparisons of all models within a realistic framework.
|
