Estimation of the Mean in a Variance Component Model - a Comparison of Estimation Procedures of ANOVA, MINQ and Sampling Theory

✍ Scribed by Dr. R. Trommer

• Publisher: John Wiley and Sons
• Year: 1985
• Tongue: English
• Weight: 475 KB
• Volume: 27
• Category: Article
• ISSN: 0323-3847
• DOI: 10.1002/bimj.4710270402

✦ Synopsis

We consider two simple estimators of the mean p in a variance component model (one-way classification, unbalanced case) 1 1 FI=; 7 Yjk; F2=; 2 g j *

I

I n dependence of the estimation procedure (ANOVA, MINQ, sampling theory) we obtain several different estimators for t h e variances of pl and pz.

I n t h e theory of two-stagesampling the sample size n = 2 ni is a random variable. Consideration of the variance component model under this assumption leads to a n interesting result. Under the supplementary assumption of vanishing covariances between some functions of nj and observations yjk we obtain the same estimators for the variances of p, and ,dz as in the common ease of constant n, and n.