Alternating logistic regressions with improved finite sample properties

Jamie Perin, John S. Preisser

Research output: Research - peer-reviewArticle

Abstract

Alternating logistic regressions is an estimating equations procedure used to model marginal means of correlated binary outcomes while simultaneously specifying a within-cluster association model for log odds ratios of outcome pairs. A recent generalization of alternating logistic regressions, known as orthogonalized residuals, is extended to incorporate finite sample adjustments in the estimation of the log odds ratio model parameters for when there is a moderately small number of clusters. Bias adjustments are made both in the sandwich variance estimators and in the estimating equations for the association parameters. The proposed methods are demonstrated in a repeated cross-sectional cluster trial to reduce underage drinking in the United States, and in an analysis of dental caries incidence in a cluster randomized trial of 30 aboriginal communities in the Northern Territory of Australia. A simulation study demonstrates improved performance with respect to bias and coverage of their estimators relative to those based on the uncorrected orthogonalized residuals procedure.

LanguageEnglish (US)
Pages696-705
Number of pages10
JournalBiometrics
Volume73
Issue number2
DOIs
StatePublished - Jun 1 2017

Fingerprint

Logistic Regression
Logistic Models
sampling
Logistics
Estimating Equation
Odds Ratio
Adjustment
Social Adjustment
odds ratio
Sandwich Estimator
Association Model
Marginal Model
Binary Outcomes
Randomized Trial
Variance Estimator
Number of Clusters
Incidence
Coverage
Simulation Study
Estimator

Keywords

  • Cluster randomized trials
  • Dental caries
  • Generalized Estimating Equations
  • Marginal association modeling
  • Small samples
  • Underage drinking

ASJC Scopus subject areas

  • Statistics and Probability
  • Biochemistry, Genetics and Molecular Biology(all)
  • Immunology and Microbiology(all)
  • Agricultural and Biological Sciences(all)
  • Applied Mathematics

Cite this

Alternating logistic regressions with improved finite sample properties. / Perin, Jamie; Preisser, John S.

In: Biometrics, Vol. 73, No. 2, 01.06.2017, p. 696-705.

Research output: Research - peer-reviewArticle

@article{fad621b605a041b79960a2655c0ef612,
title = "Alternating logistic regressions with improved finite sample properties",
abstract = "Alternating logistic regressions is an estimating equations procedure used to model marginal means of correlated binary outcomes while simultaneously specifying a within-cluster association model for log odds ratios of outcome pairs. A recent generalization of alternating logistic regressions, known as orthogonalized residuals, is extended to incorporate finite sample adjustments in the estimation of the log odds ratio model parameters for when there is a moderately small number of clusters. Bias adjustments are made both in the sandwich variance estimators and in the estimating equations for the association parameters. The proposed methods are demonstrated in a repeated cross-sectional cluster trial to reduce underage drinking in the United States, and in an analysis of dental caries incidence in a cluster randomized trial of 30 aboriginal communities in the Northern Territory of Australia. A simulation study demonstrates improved performance with respect to bias and coverage of their estimators relative to those based on the uncorrected orthogonalized residuals procedure.",
keywords = "Cluster randomized trials, Dental caries, Generalized Estimating Equations, Marginal association modeling, Small samples, Underage drinking",
author = "Jamie Perin and Preisser, {John S.}",
year = "2017",
month = "6",
doi = "10.1111/biom.12614",
volume = "73",
pages = "696--705",
journal = "Biometrics",
issn = "0006-341X",
publisher = "Wiley-Blackwell",
number = "2",

}

TY - JOUR

T1 - Alternating logistic regressions with improved finite sample properties

AU - Perin,Jamie

AU - Preisser,John S.

PY - 2017/6/1

Y1 - 2017/6/1

N2 - Alternating logistic regressions is an estimating equations procedure used to model marginal means of correlated binary outcomes while simultaneously specifying a within-cluster association model for log odds ratios of outcome pairs. A recent generalization of alternating logistic regressions, known as orthogonalized residuals, is extended to incorporate finite sample adjustments in the estimation of the log odds ratio model parameters for when there is a moderately small number of clusters. Bias adjustments are made both in the sandwich variance estimators and in the estimating equations for the association parameters. The proposed methods are demonstrated in a repeated cross-sectional cluster trial to reduce underage drinking in the United States, and in an analysis of dental caries incidence in a cluster randomized trial of 30 aboriginal communities in the Northern Territory of Australia. A simulation study demonstrates improved performance with respect to bias and coverage of their estimators relative to those based on the uncorrected orthogonalized residuals procedure.

AB - Alternating logistic regressions is an estimating equations procedure used to model marginal means of correlated binary outcomes while simultaneously specifying a within-cluster association model for log odds ratios of outcome pairs. A recent generalization of alternating logistic regressions, known as orthogonalized residuals, is extended to incorporate finite sample adjustments in the estimation of the log odds ratio model parameters for when there is a moderately small number of clusters. Bias adjustments are made both in the sandwich variance estimators and in the estimating equations for the association parameters. The proposed methods are demonstrated in a repeated cross-sectional cluster trial to reduce underage drinking in the United States, and in an analysis of dental caries incidence in a cluster randomized trial of 30 aboriginal communities in the Northern Territory of Australia. A simulation study demonstrates improved performance with respect to bias and coverage of their estimators relative to those based on the uncorrected orthogonalized residuals procedure.

KW - Cluster randomized trials

KW - Dental caries

KW - Generalized Estimating Equations

KW - Marginal association modeling

KW - Small samples

KW - Underage drinking

UR - http://www.scopus.com/inward/record.url?scp=84995804170&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84995804170&partnerID=8YFLogxK

U2 - 10.1111/biom.12614

DO - 10.1111/biom.12614

M3 - Article

VL - 73

SP - 696

EP - 705

JO - Biometrics

T2 - Biometrics

JF - Biometrics

SN - 0006-341X

IS - 2

ER -