Adaptive interval estimation in one-way random effects models

M. Y. El-Bassiouni, Taoufik Zoubeidi

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

Two-stage procedures are introduced to control the width and coverage (validity) of confidence intervals for the estimation of the mean, the between groups variance component and certain ratios of the variance components in one-way random effects models. The procedures use the pilot sample data to estimate an "optimal" group size and then proceed to determine the number of groups by a stopping rule. Such sampling plans give rise to unbalanced data, which are consequently analyzed by the harmonic mean method. Several asymptotic results concerning the proposed procedures are given along with simulation results to assess their performance in moderate sample size situations. The proposed procedures were found to effectively control the width and probability of coverage of the resulting confidence intervals in all cases and were also found to be robust in the presence of missing observations. From a practical point of view, the procedures are illustrated using a real data set and it is shown that the resulting unbalanced designs tend to require smaller sample sizes than is needed in a corresponding balanced design where the group size is arbitrarily pre-specified.

Original languageEnglish
Pages (from-to)857-872
Number of pages16
JournalJournal of Statistical Planning and Inference
Volume138
Issue number4
DOIs
Publication statusPublished - Apr 1 2008

Keywords

  • Confidence intervals
  • Harmonic mean method
  • Sequential estimation
  • Stopping time
  • Two-stage sampling
  • Validity of confidence intervals

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Adaptive interval estimation in one-way random effects models'. Together they form a unique fingerprint.

Cite this