Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/192590 
Year of Publication: 
2010
Series/Report no.: 
Discussion Papers No. 608
Publisher: 
Statistics Norway, Research Department, Oslo
Abstract: 
The design-approach is evaluated, using a likelihood approach to survey sampling. It is argued that a model-based approach is unavoidable from a scientific point of view. Estimating population quantities can then be regarded as a prediction problem. Predictive likelihood methods are considered in various cases, and evaluated by properties of related confidence intervals and asymptotic consistency.
Subjects: 
likelihood
predictive likelihood
asymptotic consistency.
JEL: 
C10
Document Type: 
Working Paper

Files in This Item:
File
Size
289.27 kB





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.