Cargando…

USP: an independence test that improves on Pearson’s chi-squared and the G-test

We present the [Formula: see text]-statistic permutation (USP) test of independence in the context of discrete data displayed in a contingency table. Either Pearson’s [Formula: see text]-test of independence, or the [Formula: see text]-test, are typically used for this task, but we argue that these...

Descripción completa

Detalles Bibliográficos
Autores principales: Berrett, Thomas B., Samworth, Richard J.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: The Royal Society 2021
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8652272/
https://www.ncbi.nlm.nih.gov/pubmed/35153605
http://dx.doi.org/10.1098/rspa.2021.0549
_version_ 1784611560631566336
author Berrett, Thomas B.
Samworth, Richard J.
author_facet Berrett, Thomas B.
Samworth, Richard J.
author_sort Berrett, Thomas B.
collection PubMed
description We present the [Formula: see text]-statistic permutation (USP) test of independence in the context of discrete data displayed in a contingency table. Either Pearson’s [Formula: see text]-test of independence, or the [Formula: see text]-test, are typically used for this task, but we argue that these tests have serious deficiencies, both in terms of their inability to control the size of the test, and their power properties. By contrast, the USP test is guaranteed to control the size of the test at the nominal level for all sample sizes, has no issues with small (or zero) cell counts, and is able to detect distributions that violate independence in only a minimal way. The test statistic is derived from a [Formula: see text]-statistic estimator of a natural population measure of dependence, and we prove that this is the unique minimum variance unbiased estimator of this population quantity. The practical utility of the USP test is demonstrated on both simulated data, where its power can be dramatically greater than those of Pearson’s test, the [Formula: see text]-test and Fisher’s exact test, and on real data. The USP test is implemented in the R package USP.
format Online
Article
Text
id pubmed-8652272
institution National Center for Biotechnology Information
language English
publishDate 2021
publisher The Royal Society
record_format MEDLINE/PubMed
spelling pubmed-86522722022-02-11 USP: an independence test that improves on Pearson’s chi-squared and the G-test Berrett, Thomas B. Samworth, Richard J. Proc Math Phys Eng Sci Research Articles We present the [Formula: see text]-statistic permutation (USP) test of independence in the context of discrete data displayed in a contingency table. Either Pearson’s [Formula: see text]-test of independence, or the [Formula: see text]-test, are typically used for this task, but we argue that these tests have serious deficiencies, both in terms of their inability to control the size of the test, and their power properties. By contrast, the USP test is guaranteed to control the size of the test at the nominal level for all sample sizes, has no issues with small (or zero) cell counts, and is able to detect distributions that violate independence in only a minimal way. The test statistic is derived from a [Formula: see text]-statistic estimator of a natural population measure of dependence, and we prove that this is the unique minimum variance unbiased estimator of this population quantity. The practical utility of the USP test is demonstrated on both simulated data, where its power can be dramatically greater than those of Pearson’s test, the [Formula: see text]-test and Fisher’s exact test, and on real data. The USP test is implemented in the R package USP. The Royal Society 2021-12 2021-12-08 /pmc/articles/PMC8652272/ /pubmed/35153605 http://dx.doi.org/10.1098/rspa.2021.0549 Text en © 2021 The Authors. https://creativecommons.org/licenses/by/4.0/Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/ (https://creativecommons.org/licenses/by/4.0/) , which permits unrestricted use, provided the original author and source are credited.
spellingShingle Research Articles
Berrett, Thomas B.
Samworth, Richard J.
USP: an independence test that improves on Pearson’s chi-squared and the G-test
title USP: an independence test that improves on Pearson’s chi-squared and the G-test
title_full USP: an independence test that improves on Pearson’s chi-squared and the G-test
title_fullStr USP: an independence test that improves on Pearson’s chi-squared and the G-test
title_full_unstemmed USP: an independence test that improves on Pearson’s chi-squared and the G-test
title_short USP: an independence test that improves on Pearson’s chi-squared and the G-test
title_sort usp: an independence test that improves on pearson’s chi-squared and the g-test
topic Research Articles
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8652272/
https://www.ncbi.nlm.nih.gov/pubmed/35153605
http://dx.doi.org/10.1098/rspa.2021.0549
work_keys_str_mv AT berrettthomasb uspanindependencetestthatimprovesonpearsonschisquaredandthegtest
AT samworthrichardj uspanindependencetestthatimprovesonpearsonschisquaredandthegtest