Statistical power of multiplicity adjustment strategies for correlated binary endpoints

Andrew C. Leon, Moonseong Heo, Jedediah J. Teres, Toshihiko Morikawa

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

There are numerous alternatives to the so-called Bonferroni adjustment to control for familywise Type I error among multiple tests. Yet, for the most part, these approaches disregard the correlation among endpoints. This can prove to be a conservative hypothesis testing strategy if the null hypothesis is false. The James procedure was proposed to account for the correlation structure among multiple continuous endpoints. Here, a simulation study evaluates the statistical power of the Hochberg and James adjustment strategies relative to that of the Bonferroni approach when used for multiple correlated binary variables. The simulations demonstrate that relative to the Bonferroni approach, neither alternative sacrifices power. The Hochberg approach has more statistical power for ρ<0.50; whereas the James procedure provides more statistical power with higher ρ, the common correlation among the multiple outcomes. A study of gender differences in New York City homicides is used to illustrate the approaches.

Original languageEnglish (US)
Pages (from-to)1712-1723
Number of pages12
JournalStatistics in Medicine
Volume26
Issue number8
DOIs
StatePublished - Apr 15 2007
Externally publishedYes

Fingerprint

Bonferroni
Statistical Power
Multiplicity
Adjustment
Binary
Homicide
Multiple Outcomes
Gender Differences
Multiple Tests
Binary Variables
Type I error
Alternatives
Correlation Structure
Hypothesis Testing
Null hypothesis
Simulation Study
Evaluate
Demonstrate
Strategy
Simulation

Keywords

  • Bonferroni adjustment
  • Correlated binary endpoints
  • James adjustment
  • Multiple comparisons
  • Multiplicity
  • Statistical power

ASJC Scopus subject areas

  • Epidemiology

Cite this

Statistical power of multiplicity adjustment strategies for correlated binary endpoints. / Leon, Andrew C.; Heo, Moonseong; Teres, Jedediah J.; Morikawa, Toshihiko.

In: Statistics in Medicine, Vol. 26, No. 8, 15.04.2007, p. 1712-1723.

Research output: Contribution to journalArticle

Leon, Andrew C. ; Heo, Moonseong ; Teres, Jedediah J. ; Morikawa, Toshihiko. / Statistical power of multiplicity adjustment strategies for correlated binary endpoints. In: Statistics in Medicine. 2007 ; Vol. 26, No. 8. pp. 1712-1723.
@article{6bee96c3485c42e9bda6126cb0bade87,
title = "Statistical power of multiplicity adjustment strategies for correlated binary endpoints",
abstract = "There are numerous alternatives to the so-called Bonferroni adjustment to control for familywise Type I error among multiple tests. Yet, for the most part, these approaches disregard the correlation among endpoints. This can prove to be a conservative hypothesis testing strategy if the null hypothesis is false. The James procedure was proposed to account for the correlation structure among multiple continuous endpoints. Here, a simulation study evaluates the statistical power of the Hochberg and James adjustment strategies relative to that of the Bonferroni approach when used for multiple correlated binary variables. The simulations demonstrate that relative to the Bonferroni approach, neither alternative sacrifices power. The Hochberg approach has more statistical power for ρ<0.50; whereas the James procedure provides more statistical power with higher ρ, the common correlation among the multiple outcomes. A study of gender differences in New York City homicides is used to illustrate the approaches.",
keywords = "Bonferroni adjustment, Correlated binary endpoints, James adjustment, Multiple comparisons, Multiplicity, Statistical power",
author = "Leon, {Andrew C.} and Moonseong Heo and Teres, {Jedediah J.} and Toshihiko Morikawa",
year = "2007",
month = "4",
day = "15",
doi = "10.1002/sim.2795",
language = "English (US)",
volume = "26",
pages = "1712--1723",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "8",

}

TY - JOUR

T1 - Statistical power of multiplicity adjustment strategies for correlated binary endpoints

AU - Leon, Andrew C.

AU - Heo, Moonseong

AU - Teres, Jedediah J.

AU - Morikawa, Toshihiko

PY - 2007/4/15

Y1 - 2007/4/15

N2 - There are numerous alternatives to the so-called Bonferroni adjustment to control for familywise Type I error among multiple tests. Yet, for the most part, these approaches disregard the correlation among endpoints. This can prove to be a conservative hypothesis testing strategy if the null hypothesis is false. The James procedure was proposed to account for the correlation structure among multiple continuous endpoints. Here, a simulation study evaluates the statistical power of the Hochberg and James adjustment strategies relative to that of the Bonferroni approach when used for multiple correlated binary variables. The simulations demonstrate that relative to the Bonferroni approach, neither alternative sacrifices power. The Hochberg approach has more statistical power for ρ<0.50; whereas the James procedure provides more statistical power with higher ρ, the common correlation among the multiple outcomes. A study of gender differences in New York City homicides is used to illustrate the approaches.

AB - There are numerous alternatives to the so-called Bonferroni adjustment to control for familywise Type I error among multiple tests. Yet, for the most part, these approaches disregard the correlation among endpoints. This can prove to be a conservative hypothesis testing strategy if the null hypothesis is false. The James procedure was proposed to account for the correlation structure among multiple continuous endpoints. Here, a simulation study evaluates the statistical power of the Hochberg and James adjustment strategies relative to that of the Bonferroni approach when used for multiple correlated binary variables. The simulations demonstrate that relative to the Bonferroni approach, neither alternative sacrifices power. The Hochberg approach has more statistical power for ρ<0.50; whereas the James procedure provides more statistical power with higher ρ, the common correlation among the multiple outcomes. A study of gender differences in New York City homicides is used to illustrate the approaches.

KW - Bonferroni adjustment

KW - Correlated binary endpoints

KW - James adjustment

KW - Multiple comparisons

KW - Multiplicity

KW - Statistical power

UR - http://www.scopus.com/inward/record.url?scp=33947383058&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33947383058&partnerID=8YFLogxK

U2 - 10.1002/sim.2795

DO - 10.1002/sim.2795

M3 - Article

C2 - 17252534

AN - SCOPUS:33947383058

VL - 26

SP - 1712

EP - 1723

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 8

ER -