Variance estimation for measures of change in probability sampling [Estimation de la variance de mesures de changement pour des échantillons probabilistes]
Variance estimation for measures of change in probability sampling [Estimation de la variance de mesures de changement pour des échantillons probabilistes]
Variance estimation of changes requires estimates of variances and covariances that would be relatively straightforward to make if the sample remained the same from one wave to the next, but this is rarely the case in practice as successive waves are usually different overlapping samples. The author proposes a design-based estimator for covariance matrices that is adapted to this situation. Under certain conditions, he shows that his approach yields non-negative definite estimates for covariance matrices and therefore positive variance estimates for a large class of measures of change.
L'estimation de la variance de changements s'appuie sur des estimés de variances et de covariances qui seraient relativement simples à obtenir si l'échantillon restait le même d'une période à l'autre, chose rare en pratique puisque les échantillons successifs sont des échantillons qui se chevauchent. L'auteur propose un estimateur de matrices de covariance adapté à cette situation. Sous certaines conditions, il montre que son approche conduit à des estimations de matrices de covariance définies non négatives et donc à des estimations de variance positives pour une grande classe de mesures de changement.
conditional poisson sampling, covariance matrix, design-based inference, entropy, finite population correction, inclusion probabilities, non-negative definite matrix, overlapping samples, repeated surveys, rotating schemes
451-467
Berger, Yves G.
8fd6af5c-31e6-4130-8b53-90910bf2f43b
December 2004
Berger, Yves G.
8fd6af5c-31e6-4130-8b53-90910bf2f43b
Berger, Yves G.
(2004)
Variance estimation for measures of change in probability sampling [Estimation de la variance de mesures de changement pour des échantillons probabilistes].
Canadian Journal of Statistics, 32 (4), .
Abstract
Variance estimation of changes requires estimates of variances and covariances that would be relatively straightforward to make if the sample remained the same from one wave to the next, but this is rarely the case in practice as successive waves are usually different overlapping samples. The author proposes a design-based estimator for covariance matrices that is adapted to this situation. Under certain conditions, he shows that his approach yields non-negative definite estimates for covariance matrices and therefore positive variance estimates for a large class of measures of change.
L'estimation de la variance de changements s'appuie sur des estimés de variances et de covariances qui seraient relativement simples à obtenir si l'échantillon restait le même d'une période à l'autre, chose rare en pratique puisque les échantillons successifs sont des échantillons qui se chevauchent. L'auteur propose un estimateur de matrices de covariance adapté à cette situation. Sous certaines conditions, il montre que son approche conduit à des estimations de matrices de covariance définies non négatives et donc à des estimations de variance positives pour une grande classe de mesures de changement.
Text
34122.pdf
- Version of Record
Restricted to Repository staff only
Request a copy
More information
Published date: December 2004
Keywords:
conditional poisson sampling, covariance matrix, design-based inference, entropy, finite population correction, inclusion probabilities, non-negative definite matrix, overlapping samples, repeated surveys, rotating schemes
Identifiers
Local EPrints ID: 34122
URI: http://eprints.soton.ac.uk/id/eprint/34122
ISSN: 0319-5724
PURE UUID: 80e95dc3-a7f9-40bc-bc90-c9c362102f57
Catalogue record
Date deposited: 17 May 2006
Last modified: 16 Mar 2024 03:03
Export record
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
