Optimal one-way random effects designs for the intraclass correlation based on confidence intervals

Brent D Burch, Ian R. Harris

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Confidence intervals for the intraclass correlation coefficient (p) are used to determine the optimal allocation of experimental material in one-way random effects models. Designs that produce narrow intervals are preferred since they provide greater precision to estimate ρ. Assuming the total cost and the relative cost of the two stages of sampling are fixed, the authors investigate the number of classes and the number of individuals per class required to minimize the expected length of confidence intervals. We obtain results using asymptotic theory and compare these results to those obtained using exact calculations. The best design depends on the unknown value of p. Minimizing the maximum expected length of confidence intervals guards against worst-case scenarios. A good overall recommendation based on asymptotic results is to choose a design having classes of size 2 + √4 + 3r, where r is the relative cost of sampling at the class-level compared to the individual-level. If r = 0, then the overall cost is the sample size and the recommendation reduces to a design having classes of size 4.

Original languageEnglish (US)
Pages (from-to)2009-2023
Number of pages15
JournalCommunications in Statistics - Theory and Methods
Volume34
Issue number9-10
DOIs
StatePublished - 2005

Fingerprint

Intraclass Correlation
Random Effects
Confidence interval
Expected Length
Costs
Sampling
Recommendations
Intraclass Correlation Coefficient
Random Effects Model
Optimal Allocation
Asymptotic Theory
Sample Size
Choose
Design
Class
Minimise
Unknown
Scenarios
Interval
Estimate

Keywords

  • Expected length
  • Optimal allocation
  • Variance components

ASJC Scopus subject areas

  • Statistics and Probability
  • Safety, Risk, Reliability and Quality

Cite this

Optimal one-way random effects designs for the intraclass correlation based on confidence intervals. / Burch, Brent D; Harris, Ian R.

In: Communications in Statistics - Theory and Methods, Vol. 34, No. 9-10, 2005, p. 2009-2023.

Research output: Contribution to journalArticle

@article{233b623d192845f89b77e9c8c1397096,
title = "Optimal one-way random effects designs for the intraclass correlation based on confidence intervals",
abstract = "Confidence intervals for the intraclass correlation coefficient (p) are used to determine the optimal allocation of experimental material in one-way random effects models. Designs that produce narrow intervals are preferred since they provide greater precision to estimate ρ. Assuming the total cost and the relative cost of the two stages of sampling are fixed, the authors investigate the number of classes and the number of individuals per class required to minimize the expected length of confidence intervals. We obtain results using asymptotic theory and compare these results to those obtained using exact calculations. The best design depends on the unknown value of p. Minimizing the maximum expected length of confidence intervals guards against worst-case scenarios. A good overall recommendation based on asymptotic results is to choose a design having classes of size 2 + √4 + 3r, where r is the relative cost of sampling at the class-level compared to the individual-level. If r = 0, then the overall cost is the sample size and the recommendation reduces to a design having classes of size 4.",
keywords = "Expected length, Optimal allocation, Variance components",
author = "Burch, {Brent D} and Harris, {Ian R.}",
year = "2005",
doi = "10.1080/03610920500203372",
language = "English (US)",
volume = "34",
pages = "2009--2023",
journal = "Communications in Statistics - Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "9-10",

}

TY - JOUR

T1 - Optimal one-way random effects designs for the intraclass correlation based on confidence intervals

AU - Burch, Brent D

AU - Harris, Ian R.

PY - 2005

Y1 - 2005

N2 - Confidence intervals for the intraclass correlation coefficient (p) are used to determine the optimal allocation of experimental material in one-way random effects models. Designs that produce narrow intervals are preferred since they provide greater precision to estimate ρ. Assuming the total cost and the relative cost of the two stages of sampling are fixed, the authors investigate the number of classes and the number of individuals per class required to minimize the expected length of confidence intervals. We obtain results using asymptotic theory and compare these results to those obtained using exact calculations. The best design depends on the unknown value of p. Minimizing the maximum expected length of confidence intervals guards against worst-case scenarios. A good overall recommendation based on asymptotic results is to choose a design having classes of size 2 + √4 + 3r, where r is the relative cost of sampling at the class-level compared to the individual-level. If r = 0, then the overall cost is the sample size and the recommendation reduces to a design having classes of size 4.

AB - Confidence intervals for the intraclass correlation coefficient (p) are used to determine the optimal allocation of experimental material in one-way random effects models. Designs that produce narrow intervals are preferred since they provide greater precision to estimate ρ. Assuming the total cost and the relative cost of the two stages of sampling are fixed, the authors investigate the number of classes and the number of individuals per class required to minimize the expected length of confidence intervals. We obtain results using asymptotic theory and compare these results to those obtained using exact calculations. The best design depends on the unknown value of p. Minimizing the maximum expected length of confidence intervals guards against worst-case scenarios. A good overall recommendation based on asymptotic results is to choose a design having classes of size 2 + √4 + 3r, where r is the relative cost of sampling at the class-level compared to the individual-level. If r = 0, then the overall cost is the sample size and the recommendation reduces to a design having classes of size 4.

KW - Expected length

KW - Optimal allocation

KW - Variance components

UR - http://www.scopus.com/inward/record.url?scp=25644450272&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=25644450272&partnerID=8YFLogxK

U2 - 10.1080/03610920500203372

DO - 10.1080/03610920500203372

M3 - Article

AN - SCOPUS:25644450272

VL - 34

SP - 2009

EP - 2023

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 9-10

ER -