The split sample permutation t-tests

Shunpu Zhang

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Without the exchangeability assumption, permutation tests for comparing two population means do not provide exact control of the probability of making a Type I error. Another drawback of permutation tests is that it cannot be used to test hypothesis about one population. In this paper, we propose a new type of permutation tests for testing the difference between two population means: the split sample permutation t-tests. We show that the split sample permutation t-tests do not require the exchangeability assumption, are asymptotically exact and can be easily extended to testing hypothesis about one population. Extensive simulations were carried out to evaluate the performance of two specific split sample permutation t-tests: the split in the middle permutation t-test and the split in the end permutation t-test. The simulation results show that the split in the middle permutation t-test has comparable performance to the permutation test if the population distributions are symmetric and satisfy the exchangeability assumption. Otherwise, the split in the end permutation t-test has significantly more accurate control of level of significance than the split in the middle permutation t-test and other existing permutation tests.

Original languageEnglish (US)
Pages (from-to)3512-3524
Number of pages13
JournalJournal of Statistical Planning and Inference
Volume139
Issue number10
DOIs
StatePublished - Oct 1 2009

Fingerprint

Permutation Test
t-test
Population distribution
Testing
Exchangeability
T-test
Testing Hypotheses
Type I error
Hypothesis Test
Simulation

Keywords

  • Exact test
  • Exchangeability
  • Permutation tests

ASJC Scopus subject areas

  • Statistics, Probability and Uncertainty
  • Applied Mathematics
  • Statistics and Probability

Cite this

The split sample permutation t-tests. / Zhang, Shunpu.

In: Journal of Statistical Planning and Inference, Vol. 139, No. 10, 01.10.2009, p. 3512-3524.

Research output: Contribution to journalArticle

Zhang, Shunpu. / The split sample permutation t-tests. In: Journal of Statistical Planning and Inference. 2009 ; Vol. 139, No. 10. pp. 3512-3524.
@article{41e75e7db7c040f0ad049d1b3b148529,
title = "The split sample permutation t-tests",
abstract = "Without the exchangeability assumption, permutation tests for comparing two population means do not provide exact control of the probability of making a Type I error. Another drawback of permutation tests is that it cannot be used to test hypothesis about one population. In this paper, we propose a new type of permutation tests for testing the difference between two population means: the split sample permutation t-tests. We show that the split sample permutation t-tests do not require the exchangeability assumption, are asymptotically exact and can be easily extended to testing hypothesis about one population. Extensive simulations were carried out to evaluate the performance of two specific split sample permutation t-tests: the split in the middle permutation t-test and the split in the end permutation t-test. The simulation results show that the split in the middle permutation t-test has comparable performance to the permutation test if the population distributions are symmetric and satisfy the exchangeability assumption. Otherwise, the split in the end permutation t-test has significantly more accurate control of level of significance than the split in the middle permutation t-test and other existing permutation tests.",
keywords = "Exact test, Exchangeability, Permutation tests",
author = "Shunpu Zhang",
year = "2009",
month = "10",
day = "1",
doi = "10.1016/j.jspi.2009.04.004",
language = "English (US)",
volume = "139",
pages = "3512--3524",
journal = "Journal of Statistical Planning and Inference",
issn = "0378-3758",
publisher = "Elsevier",
number = "10",

}

TY - JOUR

T1 - The split sample permutation t-tests

AU - Zhang, Shunpu

PY - 2009/10/1

Y1 - 2009/10/1

N2 - Without the exchangeability assumption, permutation tests for comparing two population means do not provide exact control of the probability of making a Type I error. Another drawback of permutation tests is that it cannot be used to test hypothesis about one population. In this paper, we propose a new type of permutation tests for testing the difference between two population means: the split sample permutation t-tests. We show that the split sample permutation t-tests do not require the exchangeability assumption, are asymptotically exact and can be easily extended to testing hypothesis about one population. Extensive simulations were carried out to evaluate the performance of two specific split sample permutation t-tests: the split in the middle permutation t-test and the split in the end permutation t-test. The simulation results show that the split in the middle permutation t-test has comparable performance to the permutation test if the population distributions are symmetric and satisfy the exchangeability assumption. Otherwise, the split in the end permutation t-test has significantly more accurate control of level of significance than the split in the middle permutation t-test and other existing permutation tests.

AB - Without the exchangeability assumption, permutation tests for comparing two population means do not provide exact control of the probability of making a Type I error. Another drawback of permutation tests is that it cannot be used to test hypothesis about one population. In this paper, we propose a new type of permutation tests for testing the difference between two population means: the split sample permutation t-tests. We show that the split sample permutation t-tests do not require the exchangeability assumption, are asymptotically exact and can be easily extended to testing hypothesis about one population. Extensive simulations were carried out to evaluate the performance of two specific split sample permutation t-tests: the split in the middle permutation t-test and the split in the end permutation t-test. The simulation results show that the split in the middle permutation t-test has comparable performance to the permutation test if the population distributions are symmetric and satisfy the exchangeability assumption. Otherwise, the split in the end permutation t-test has significantly more accurate control of level of significance than the split in the middle permutation t-test and other existing permutation tests.

KW - Exact test

KW - Exchangeability

KW - Permutation tests

UR - http://www.scopus.com/inward/record.url?scp=67650001639&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67650001639&partnerID=8YFLogxK

U2 - 10.1016/j.jspi.2009.04.004

DO - 10.1016/j.jspi.2009.04.004

M3 - Article

VL - 139

SP - 3512

EP - 3524

JO - Journal of Statistical Planning and Inference

JF - Journal of Statistical Planning and Inference

SN - 0378-3758

IS - 10

ER -