Combining p-values via averaging. / Vovk, Vladimir; Wang, Ruodu.

In: Biometrika, Vol. 107, No. 4, asaa027, 11.06.2020, p. 791-808.

Research output: Contribution to journalArticlepeer-review

Published

Standard

Combining p-values via averaging. / Vovk, Vladimir; Wang, Ruodu.

In: Biometrika, Vol. 107, No. 4, asaa027, 11.06.2020, p. 791-808.

Research output: Contribution to journalArticlepeer-review

Harvard

Vovk, V & Wang, R 2020, 'Combining p-values via averaging', Biometrika, vol. 107, no. 4, asaa027, pp. 791-808. https://doi.org/10.1093/biomet/asaa027

APA

Vovk, V., & Wang, R. (2020). Combining p-values via averaging. Biometrika, 107(4), 791-808. [asaa027]. https://doi.org/10.1093/biomet/asaa027

Vancouver

Vovk V, Wang R. Combining p-values via averaging. Biometrika. 2020 Jun 11;107(4):791-808. asaa027. https://doi.org/10.1093/biomet/asaa027

Author

Vovk, Vladimir ; Wang, Ruodu. / Combining p-values via averaging. In: Biometrika. 2020 ; Vol. 107, No. 4. pp. 791-808.

BibTeX

@article{3285831e2029452cbf7a8cace5c2739d,
title = "Combining p-values via averaging",
abstract = "This paper proposes general methods for the problem of multiple testing of a single hypothesis, with a standard goal of combining a number of p-values without making any assumptions about their dependence structure. An old result by Ruschendorf and, independently, Meng implies that the p-values can be combined by scaling up their arithmetic mean by a factor of 2, and no smaller factor is sufficient in general. A similar result by Mattner about the geometric mean replaces 2 by e. Based on more recent developments in mathematical finance, specifically, robust risk aggregation techniques, we extend these results to generalized means; in particular, we show that K p-values can be combined by scaling up their harmonic mean by a factor of \log K asymptotically as K\to\infty. This leads to a generalized version of the Bonferroni-Holm procedure. We also explore methods using weighted averages of p-values. Finally, we discuss the efficiency of various methods of combining p-values and how to choose a suitable method in light of data and prior information.",
keywords = "Hypothesis testing, Multiple hypothesis testing, Multiple testing of a single hypothesis, Robust risk aggregation",
author = "Vladimir Vovk and Ruodu Wang",
year = "2020",
month = jun,
day = "11",
doi = "10.1093/biomet/asaa027",
language = "English",
volume = "107",
pages = "791--808",
journal = "Biometrika",
issn = "0006-3444",
publisher = "Oxford University Press",
number = "4",

}

RIS

TY - JOUR

T1 - Combining p-values via averaging

AU - Vovk, Vladimir

AU - Wang, Ruodu

PY - 2020/6/11

Y1 - 2020/6/11

N2 - This paper proposes general methods for the problem of multiple testing of a single hypothesis, with a standard goal of combining a number of p-values without making any assumptions about their dependence structure. An old result by Ruschendorf and, independently, Meng implies that the p-values can be combined by scaling up their arithmetic mean by a factor of 2, and no smaller factor is sufficient in general. A similar result by Mattner about the geometric mean replaces 2 by e. Based on more recent developments in mathematical finance, specifically, robust risk aggregation techniques, we extend these results to generalized means; in particular, we show that K p-values can be combined by scaling up their harmonic mean by a factor of \log K asymptotically as K\to\infty. This leads to a generalized version of the Bonferroni-Holm procedure. We also explore methods using weighted averages of p-values. Finally, we discuss the efficiency of various methods of combining p-values and how to choose a suitable method in light of data and prior information.

AB - This paper proposes general methods for the problem of multiple testing of a single hypothesis, with a standard goal of combining a number of p-values without making any assumptions about their dependence structure. An old result by Ruschendorf and, independently, Meng implies that the p-values can be combined by scaling up their arithmetic mean by a factor of 2, and no smaller factor is sufficient in general. A similar result by Mattner about the geometric mean replaces 2 by e. Based on more recent developments in mathematical finance, specifically, robust risk aggregation techniques, we extend these results to generalized means; in particular, we show that K p-values can be combined by scaling up their harmonic mean by a factor of \log K asymptotically as K\to\infty. This leads to a generalized version of the Bonferroni-Holm procedure. We also explore methods using weighted averages of p-values. Finally, we discuss the efficiency of various methods of combining p-values and how to choose a suitable method in light of data and prior information.

KW - Hypothesis testing

KW - Multiple hypothesis testing

KW - Multiple testing of a single hypothesis

KW - Robust risk aggregation

U2 - 10.1093/biomet/asaa027

DO - 10.1093/biomet/asaa027

M3 - Article

VL - 107

SP - 791

EP - 808

JO - Biometrika

JF - Biometrika

SN - 0006-3444

IS - 4

M1 - asaa027

ER -