Premium
Estimation of a finite universe total when a stratum is not sampled
Author(s) -
Wright Tommy
Publication year - 1999
Publication title -
applied stochastic models in business and industry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.413
H-Index - 40
eISSN - 1526-4025
pISSN - 1524-1904
DOI - 10.1002/(sici)1526-4025(199904/06)15:2<131::aid-asmb365>3.0.co;2-h
Subject(s) - context (archaeology) , sample (material) , universe , sampling (signal processing) , listing (finance) , estimator , statistics , generalization , econometrics , computer science , mathematics , geography , economics , archaeology , filter (signal processing) , astrophysics , physics , mathematical analysis , finance , computer vision , thermodynamics
In the context of a universe of trucks operating in the United States in 1990, this paper presents statistical methodology for estimating a finite universe total on a second occasion when a part of the universe is sampled and the remainder of the universe is not sampled. Prediction is used to compensate for the lack of data from the unsampled portion of the universe. The sample, stratified by age, is from an earlier census without updating the listing (frame). Accounting for births and deaths in the universe between the two points in time, an estimator is obtained which is a generalization of what an analyst might do in the absence of sample data from a given stratum, the births. Deaths are accounted for through domain estimation, and total updated counts are available from different sources. The approach of the paper is to provide an estimate for births, without actually sampling from the births stratum. With regard to saving resources by not sampling births, it is demonstrated that the analyst who does not sample the births may very well do better than the analyst who does. Copyright © 1999 John Wiley & Sons, Ltd.