Covariances and correlations in finite population sampling / Jack Gambino. : CS11-619/2013-7E-PDF

"Under the design-based framework, the derivation of covariance formulas can be tedious, but not much more tedious than the derivation of variances, which are a special case. in this paper we derive covariances under several simple sample designs, with an emphasis on partially overlapping samples. For simple random sampling, we also compare the design-based covariance to that obtained under the model-based approach. We also briefly look at correlations. The main benefit of the paper may be that it puts into one document several scattered results that are useful to survey statisticians"--Abstract.

Lien permanent pour cette publication :
publications.gc.ca/pub?id=9.852584&sl=1

• Format MARC XML
• Format MARC HTML

Demander des formats alternatifs