Investigation of the Delete-a-Group Jackknife Variance Estimator
The aim of this work was to investigate whether the delete-a-group jackknife variance estimator could be used to produce sample errors for the Household Labour Force Survey (HLFS).
The delete-a-group jackknife variance estimator is described in Kott (1998). This variance estimation method differs from the normal jackknife in that the sample is systematically divided into a relatively small number of groups, and each jackknife replicate is then formed by deleting an entire group. In the normal jackknife a replicate is formed by deleting a single sample unit, resulting in a very large number of replicates. The normal jackknife variance estimate can therefore take an unacceptably long time to run.
The method also differs from the grouped jackknife methods which are discussed in Särndal, Swensson, and Wretman (1992) in that the groups are formed across strata. This means that the number of strata in the design does not necessarily limit how few groups are used.
(1246Kb)
If you do not have the Adobe Acrobat Reader you may download the reader to view or print the contents of this file.
back to top
For enquiries and information requests
This page last modified on: 09 January 2002