|
Organizers |
Almost balanced replicate variance estimation
by
Bill Gross
Australian Bureau of Statistics
This talk describes general results for replicate variance estimators for stratified samples and uses them to derive a general method for correcting bias and to justify 'almost balanced' replication. The method of correcting bias generalises existing scaling adjustments. There are a number of methods published for creating exactly balanced sets of replicates to reduce the variance of the replicate estimator, but they are for quite restricted combinations of sample sizes, number of strata and number of replicates.
'Almost balanced' sets of replicates can be generated for any combination of sample sizes, number of strata and number of replicates.
Date received: April 21, 2002
Copyright © 2002 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # cajg-56.