Testing for conditional multiple marginal independence

Christopher R Bilder, T. M. Loughin

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Survey respondents are often prompted to pick any number of responses from a set of possible responses. Categorical variables that summarize this kind of data are called pick any/c variables. Counts from surveys that contain a pick any/c variable along with a group variable (r levels) and stratification variable (q levels) can be marginally summarized into an r × c × q contingency table. A question that may naturally arise from this setup is to determine if the group and pick any/c variable are marginally independent given the stratification variable. A test for conditional multiple marginal independence (CMMI) can be used to answer this question. Since subjects may pick any number out of c possible responses, the Cochran (1954, Biometrics 10, 417-451) and Mantel and Haenszel (1959, Journal of the National Cancer Institute 22, 719-748) tests cannot be used directly because they assume that units in the contingency table are independent of each other. Therefore, new testing methods are developed. Cochran's test statistic is extended to r × 2 × q tables, and a modified version of this statistic is proposed to test CMMI. Its sampling distribution can be approximated through bootstrapping. Other CMMI testing methods discussed are bootstrap p-value combination methods and Bonferroni adjustments. Simulation findings suggest that the proposed bootstrap procedures and the Bonferroni adjustments consistently hold the correct size and provide power against various alternatives.

Original languageEnglish (US)
Pages (from-to)200-208
Number of pages9
JournalBiometrics
Volume58
Issue number1
DOIs
StatePublished - Jan 1 2002

Fingerprint

Statistics
Testing
Biometrics
Bonferroni
testing
Contingency Table
Sampling
Stratification
Bootstrap
National Cancer Institute (U.S.)
Adjustment
statistics
Conditional Test
Categorical variable
Sampling Distribution
Bootstrapping
p-Value
biometry
Test Statistic
Statistic

Keywords

  • Bootstrap
  • Cochran test
  • Conditional independence
  • Correlated binary data
  • Mantel-Haenszel test
  • Pick any/c

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

Testing for conditional multiple marginal independence. / Bilder, Christopher R; Loughin, T. M.

In: Biometrics, Vol. 58, No. 1, 01.01.2002, p. 200-208.

Research output: Contribution to journalArticle

Bilder, Christopher R ; Loughin, T. M. / Testing for conditional multiple marginal independence. In: Biometrics. 2002 ; Vol. 58, No. 1. pp. 200-208.
@article{85a35784281b4ac3bcaba210121f959b,
title = "Testing for conditional multiple marginal independence",
abstract = "Survey respondents are often prompted to pick any number of responses from a set of possible responses. Categorical variables that summarize this kind of data are called pick any/c variables. Counts from surveys that contain a pick any/c variable along with a group variable (r levels) and stratification variable (q levels) can be marginally summarized into an r × c × q contingency table. A question that may naturally arise from this setup is to determine if the group and pick any/c variable are marginally independent given the stratification variable. A test for conditional multiple marginal independence (CMMI) can be used to answer this question. Since subjects may pick any number out of c possible responses, the Cochran (1954, Biometrics 10, 417-451) and Mantel and Haenszel (1959, Journal of the National Cancer Institute 22, 719-748) tests cannot be used directly because they assume that units in the contingency table are independent of each other. Therefore, new testing methods are developed. Cochran's test statistic is extended to r × 2 × q tables, and a modified version of this statistic is proposed to test CMMI. Its sampling distribution can be approximated through bootstrapping. Other CMMI testing methods discussed are bootstrap p-value combination methods and Bonferroni adjustments. Simulation findings suggest that the proposed bootstrap procedures and the Bonferroni adjustments consistently hold the correct size and provide power against various alternatives.",
keywords = "Bootstrap, Cochran test, Conditional independence, Correlated binary data, Mantel-Haenszel test, Pick any/c",
author = "Bilder, {Christopher R} and Loughin, {T. M.}",
year = "2002",
month = "1",
day = "1",
doi = "10.1111/j.0006-341X.2002.00200.x",
language = "English (US)",
volume = "58",
pages = "200--208",
journal = "Biometrics",
issn = "0006-341X",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - Testing for conditional multiple marginal independence

AU - Bilder, Christopher R

AU - Loughin, T. M.

PY - 2002/1/1

Y1 - 2002/1/1

N2 - Survey respondents are often prompted to pick any number of responses from a set of possible responses. Categorical variables that summarize this kind of data are called pick any/c variables. Counts from surveys that contain a pick any/c variable along with a group variable (r levels) and stratification variable (q levels) can be marginally summarized into an r × c × q contingency table. A question that may naturally arise from this setup is to determine if the group and pick any/c variable are marginally independent given the stratification variable. A test for conditional multiple marginal independence (CMMI) can be used to answer this question. Since subjects may pick any number out of c possible responses, the Cochran (1954, Biometrics 10, 417-451) and Mantel and Haenszel (1959, Journal of the National Cancer Institute 22, 719-748) tests cannot be used directly because they assume that units in the contingency table are independent of each other. Therefore, new testing methods are developed. Cochran's test statistic is extended to r × 2 × q tables, and a modified version of this statistic is proposed to test CMMI. Its sampling distribution can be approximated through bootstrapping. Other CMMI testing methods discussed are bootstrap p-value combination methods and Bonferroni adjustments. Simulation findings suggest that the proposed bootstrap procedures and the Bonferroni adjustments consistently hold the correct size and provide power against various alternatives.

AB - Survey respondents are often prompted to pick any number of responses from a set of possible responses. Categorical variables that summarize this kind of data are called pick any/c variables. Counts from surveys that contain a pick any/c variable along with a group variable (r levels) and stratification variable (q levels) can be marginally summarized into an r × c × q contingency table. A question that may naturally arise from this setup is to determine if the group and pick any/c variable are marginally independent given the stratification variable. A test for conditional multiple marginal independence (CMMI) can be used to answer this question. Since subjects may pick any number out of c possible responses, the Cochran (1954, Biometrics 10, 417-451) and Mantel and Haenszel (1959, Journal of the National Cancer Institute 22, 719-748) tests cannot be used directly because they assume that units in the contingency table are independent of each other. Therefore, new testing methods are developed. Cochran's test statistic is extended to r × 2 × q tables, and a modified version of this statistic is proposed to test CMMI. Its sampling distribution can be approximated through bootstrapping. Other CMMI testing methods discussed are bootstrap p-value combination methods and Bonferroni adjustments. Simulation findings suggest that the proposed bootstrap procedures and the Bonferroni adjustments consistently hold the correct size and provide power against various alternatives.

KW - Bootstrap

KW - Cochran test

KW - Conditional independence

KW - Correlated binary data

KW - Mantel-Haenszel test

KW - Pick any/c

UR - http://www.scopus.com/inward/record.url?scp=0036188446&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036188446&partnerID=8YFLogxK

U2 - 10.1111/j.0006-341X.2002.00200.x

DO - 10.1111/j.0006-341X.2002.00200.x

M3 - Article

VL - 58

SP - 200

EP - 208

JO - Biometrics

JF - Biometrics

SN - 0006-341X

IS - 1

ER -