GEE analysis of negatively correlated binary responses

A caution

James A. Hanley, Abdissa Negassa, Michael D DeB Edwardes

Research output: Contribution to journalArticle

22 Citations (Scopus)

Abstract

The method of generalized estimating equations has become almost standard for analysing longitudinal and other correlated response data. However, we have found that if binary responses have less than binomial variation over clusters, and are modelled using exchangeable correlations, prevailing software implementations may give unreliable results. Bounding the negative correlation away from its theoretical minimum may not always be a satisfactory solution. In such instances, using the independence working correlation structure and robust SEs is a more trustworthy alternative. Copyright (C) 2000 John Wiley and Sons, Ltd.

Original languageEnglish (US)
Pages (from-to)715-722
Number of pages8
JournalStatistics in Medicine
Volume19
Issue number5
DOIs
StatePublished - Mar 15 2000
Externally publishedYes

Fingerprint

Binary Response
Software
Generalized Estimating Equations
Correlation Structure
Alternatives
Independence
Standards

ASJC Scopus subject areas

  • Epidemiology

Cite this

GEE analysis of negatively correlated binary responses : A caution. / Hanley, James A.; Negassa, Abdissa; Edwardes, Michael D DeB.

In: Statistics in Medicine, Vol. 19, No. 5, 15.03.2000, p. 715-722.

Research output: Contribution to journalArticle

Hanley, James A. ; Negassa, Abdissa ; Edwardes, Michael D DeB. / GEE analysis of negatively correlated binary responses : A caution. In: Statistics in Medicine. 2000 ; Vol. 19, No. 5. pp. 715-722.
@article{1ccbc8e93e654e86b57fa4bf6b9ec204,
title = "GEE analysis of negatively correlated binary responses: A caution",
abstract = "The method of generalized estimating equations has become almost standard for analysing longitudinal and other correlated response data. However, we have found that if binary responses have less than binomial variation over clusters, and are modelled using exchangeable correlations, prevailing software implementations may give unreliable results. Bounding the negative correlation away from its theoretical minimum may not always be a satisfactory solution. In such instances, using the independence working correlation structure and robust SEs is a more trustworthy alternative. Copyright (C) 2000 John Wiley and Sons, Ltd.",
author = "Hanley, {James A.} and Abdissa Negassa and Edwardes, {Michael D DeB}",
year = "2000",
month = "3",
day = "15",
doi = "10.1002/(SICI)1097-0258(20000315)19:5<715::AID-SIM342>3.0.CO;2-T",
language = "English (US)",
volume = "19",
pages = "715--722",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "5",

}

TY - JOUR

T1 - GEE analysis of negatively correlated binary responses

T2 - A caution

AU - Hanley, James A.

AU - Negassa, Abdissa

AU - Edwardes, Michael D DeB

PY - 2000/3/15

Y1 - 2000/3/15

N2 - The method of generalized estimating equations has become almost standard for analysing longitudinal and other correlated response data. However, we have found that if binary responses have less than binomial variation over clusters, and are modelled using exchangeable correlations, prevailing software implementations may give unreliable results. Bounding the negative correlation away from its theoretical minimum may not always be a satisfactory solution. In such instances, using the independence working correlation structure and robust SEs is a more trustworthy alternative. Copyright (C) 2000 John Wiley and Sons, Ltd.

AB - The method of generalized estimating equations has become almost standard for analysing longitudinal and other correlated response data. However, we have found that if binary responses have less than binomial variation over clusters, and are modelled using exchangeable correlations, prevailing software implementations may give unreliable results. Bounding the negative correlation away from its theoretical minimum may not always be a satisfactory solution. In such instances, using the independence working correlation structure and robust SEs is a more trustworthy alternative. Copyright (C) 2000 John Wiley and Sons, Ltd.

UR - http://www.scopus.com/inward/record.url?scp=0034653340&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034653340&partnerID=8YFLogxK

U2 - 10.1002/(SICI)1097-0258(20000315)19:5<715::AID-SIM342>3.0.CO;2-T

DO - 10.1002/(SICI)1097-0258(20000315)19:5<715::AID-SIM342>3.0.CO;2-T

M3 - Article

VL - 19

SP - 715

EP - 722

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 5

ER -