Computer implementation of two approaches to estimating the variance components in a general p-stage staggered nested design is provided through a FORTRAN program. The first approach, the ANOVA method, is based on expected mean squares and results in unbiased estimators. The second approach, the principal component method, gives estimates which are biased but nonnegative and which have been found to be superior to the usual ANOVA estimates under various comparison criteria.
No takes yet. Share an insight, caveat, or question.
Naik et al. (1998) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: