Exact unconditional tests for a 2 × 2 matched-pairs design

Roger L. Berger, K. Sidik

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

The problem of comparing two proportions in a 2 × 2 matched-pairs design with binary responses is considered. We consider one-sided null and alternative hypotheses. The problem has two nuisance parameters. Using the monotonicity of the multinomial distribution, four exact unconditional tests based on p-values are proposed by reducing the dimension of the nuisance parameter space from two to one in computation. The size and power of the four exact tests and two other tests, the exact conditional binomial test and the asymptotic McNemar's test, are considered. It is shown that the tests based on the confidence interval p-value are more powerful than the tests based on the standard p-value. In addition, it is found that the exact conditional binomial test is conservative and not powerful for testing the hypothesis. Moreover, the asymptotic McNemar's test is shown to have incorrect size; that is, its size is larger than the nominal level of the test. Overall, the test based on McNemar's statistic and the confidence interval p-value is found to be the most powerful test with the correct size among the tests in this comparison.

Original languageEnglish (US)
Pages (from-to)91-108
Number of pages18
JournalStatistical Methods in Medical Research
Volume12
Issue number2
DOIs
StatePublished - Mar 2003
Externally publishedYes

Fingerprint

Unconditional Test
Matched pairs
Exact Test
Confidence Intervals
p-Value
McNemar's Test
Asymptotic Test
Nuisance Parameter
Confidence interval
Design
Multinomial Distribution
Binary Response
Categorical or nominal
Statistic
Null
Monotonicity
Parameter Space
Two Parameters
Proportion

ASJC Scopus subject areas

  • Epidemiology
  • Health Information Management
  • Nursing(all)

Cite this

Exact unconditional tests for a 2 × 2 matched-pairs design. / Berger, Roger L.; Sidik, K.

In: Statistical Methods in Medical Research, Vol. 12, No. 2, 03.2003, p. 91-108.

Research output: Contribution to journalArticle

Berger, Roger L. ; Sidik, K. / Exact unconditional tests for a 2 × 2 matched-pairs design. In: Statistical Methods in Medical Research. 2003 ; Vol. 12, No. 2. pp. 91-108.
@article{16f2c45d68574df9a3c5d944a502e471,
title = "Exact unconditional tests for a 2 × 2 matched-pairs design",
abstract = "The problem of comparing two proportions in a 2 × 2 matched-pairs design with binary responses is considered. We consider one-sided null and alternative hypotheses. The problem has two nuisance parameters. Using the monotonicity of the multinomial distribution, four exact unconditional tests based on p-values are proposed by reducing the dimension of the nuisance parameter space from two to one in computation. The size and power of the four exact tests and two other tests, the exact conditional binomial test and the asymptotic McNemar's test, are considered. It is shown that the tests based on the confidence interval p-value are more powerful than the tests based on the standard p-value. In addition, it is found that the exact conditional binomial test is conservative and not powerful for testing the hypothesis. Moreover, the asymptotic McNemar's test is shown to have incorrect size; that is, its size is larger than the nominal level of the test. Overall, the test based on McNemar's statistic and the confidence interval p-value is found to be the most powerful test with the correct size among the tests in this comparison.",
author = "Berger, {Roger L.} and K. Sidik",
year = "2003",
month = "3",
doi = "10.1191/0962280203sm312ra",
language = "English (US)",
volume = "12",
pages = "91--108",
journal = "Statistical Methods in Medical Research",
issn = "0962-2802",
publisher = "SAGE Publications Ltd",
number = "2",

}

TY - JOUR

T1 - Exact unconditional tests for a 2 × 2 matched-pairs design

AU - Berger, Roger L.

AU - Sidik, K.

PY - 2003/3

Y1 - 2003/3

N2 - The problem of comparing two proportions in a 2 × 2 matched-pairs design with binary responses is considered. We consider one-sided null and alternative hypotheses. The problem has two nuisance parameters. Using the monotonicity of the multinomial distribution, four exact unconditional tests based on p-values are proposed by reducing the dimension of the nuisance parameter space from two to one in computation. The size and power of the four exact tests and two other tests, the exact conditional binomial test and the asymptotic McNemar's test, are considered. It is shown that the tests based on the confidence interval p-value are more powerful than the tests based on the standard p-value. In addition, it is found that the exact conditional binomial test is conservative and not powerful for testing the hypothesis. Moreover, the asymptotic McNemar's test is shown to have incorrect size; that is, its size is larger than the nominal level of the test. Overall, the test based on McNemar's statistic and the confidence interval p-value is found to be the most powerful test with the correct size among the tests in this comparison.

AB - The problem of comparing two proportions in a 2 × 2 matched-pairs design with binary responses is considered. We consider one-sided null and alternative hypotheses. The problem has two nuisance parameters. Using the monotonicity of the multinomial distribution, four exact unconditional tests based on p-values are proposed by reducing the dimension of the nuisance parameter space from two to one in computation. The size and power of the four exact tests and two other tests, the exact conditional binomial test and the asymptotic McNemar's test, are considered. It is shown that the tests based on the confidence interval p-value are more powerful than the tests based on the standard p-value. In addition, it is found that the exact conditional binomial test is conservative and not powerful for testing the hypothesis. Moreover, the asymptotic McNemar's test is shown to have incorrect size; that is, its size is larger than the nominal level of the test. Overall, the test based on McNemar's statistic and the confidence interval p-value is found to be the most powerful test with the correct size among the tests in this comparison.

UR - http://www.scopus.com/inward/record.url?scp=0037343359&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037343359&partnerID=8YFLogxK

U2 - 10.1191/0962280203sm312ra

DO - 10.1191/0962280203sm312ra

M3 - Article

VL - 12

SP - 91

EP - 108

JO - Statistical Methods in Medical Research

JF - Statistical Methods in Medical Research

SN - 0962-2802

IS - 2

ER -