A comparison of two variance estimation techniques : the jackknife method and the linearized jackknife method / Roanna M. Beebakhee. : CS11-619/95-5E-PDF

"The jackknife method of variance estimation is used extensively for surveys with complex designs. Although it has desirable statistical properties, it is computer intensive resulting in relatively high costs. In addition, when it is used for estimating the variance of regression estimators under a stratified multi-stage cluster design, it requires the calculation of matrix inverses which has the potential of generating computational degeneracies. An alternative method, known as the linearized jackknife method has been proposed and its theoretical properties have been investigated by Yung and Rao (1995). An empirical study has been undertaken which applies both variance methods to a number of household surveys conducted by Statistics Canada. Further, the linearized jackknife variance method and the customary Taylor linearization variance method are compared to the jackknife method"--Abstract.

Lien permanent pour cette publication :
publications.gc.ca/pub?id=9.841015&sl=1

• Format MARC XML
• Format MARC HTML

Demander des formats alternatifs