Home > Research > Publications & Outputs > Four reasons to prefer Bayesian analyses over s...

Associated organisational units

Electronic data

Links

Text available via DOI:

View graph of relations

Four reasons to prefer Bayesian analyses over significance testing

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Four reasons to prefer Bayesian analyses over significance testing. / Dienes, Zoltan; McLatchie, Neil Marvin.
In: Psychonomic Bulletin and Review, Vol. 25, No. 1, 02.2018, p. 207-218.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Dienes, Z & McLatchie, NM 2018, 'Four reasons to prefer Bayesian analyses over significance testing', Psychonomic Bulletin and Review, vol. 25, no. 1, pp. 207-218. https://doi.org/10.3758/s13423-017-1266-z

APA

Vancouver

Dienes Z, McLatchie NM. Four reasons to prefer Bayesian analyses over significance testing. Psychonomic Bulletin and Review. 2018 Feb;25(1):207-218. Epub 2017 Mar 28. doi: 10.3758/s13423-017-1266-z

Author

Dienes, Zoltan ; McLatchie, Neil Marvin. / Four reasons to prefer Bayesian analyses over significance testing. In: Psychonomic Bulletin and Review. 2018 ; Vol. 25, No. 1. pp. 207-218.

Bibtex

@article{5cd534b814b4433983ef7f8b4503d511,
title = "Four reasons to prefer Bayesian analyses over significance testing",
abstract = "Inference using significance testing and Bayes factors is compared and contrasted in five case studies based on real research. The first study illustrates that the methods will often agree, both in motivating researchers to conclude that H1 is supported better than H0 and the other way round, that H0 is better supported than H1. The next four however, show that the methods will also often disagree. In these cases, the aim of the paper will be to motivate the sensible evidential conclusion and then see which approach matches those intuitions. Specifically, it is shown that a high-powered non-significant result is consistent with no evidence for H0 over H1 worth mentioning, which a Bayes factor can show; and conversely that a low-powered non-significant result is consistent with substantial evidence for H0 over H1, again indicated by Bayesian analyses. The fourth study illustrates that a high-powered significant result may not amount to any evidence for H1 over H0, matching the Bayesian conclusion. Finally the fifth study illustrates that different theories can be evidentially supported to different degrees by the same data, a fact that p-values cannot reflect but Bayes factors can. It is argued that appropriate conclusions match the Bayesian inferences, but not those based on significance testing, where they disagree.",
keywords = "Bayesian analysis, Bayes factors, Replication, Bayes factor, Bayesian statistics , Power, Significance testing , Statistics ",
author = "Zoltan Dienes and McLatchie, {Neil Marvin}",
year = "2018",
month = feb,
doi = "10.3758/s13423-017-1266-z",
language = "English",
volume = "25",
pages = "207--218",
journal = "Psychonomic Bulletin and Review",
issn = "1069-9384",
publisher = "Springer",
number = "1",

}

RIS

TY - JOUR

T1 - Four reasons to prefer Bayesian analyses over significance testing

AU - Dienes, Zoltan

AU - McLatchie, Neil Marvin

PY - 2018/2

Y1 - 2018/2

N2 - Inference using significance testing and Bayes factors is compared and contrasted in five case studies based on real research. The first study illustrates that the methods will often agree, both in motivating researchers to conclude that H1 is supported better than H0 and the other way round, that H0 is better supported than H1. The next four however, show that the methods will also often disagree. In these cases, the aim of the paper will be to motivate the sensible evidential conclusion and then see which approach matches those intuitions. Specifically, it is shown that a high-powered non-significant result is consistent with no evidence for H0 over H1 worth mentioning, which a Bayes factor can show; and conversely that a low-powered non-significant result is consistent with substantial evidence for H0 over H1, again indicated by Bayesian analyses. The fourth study illustrates that a high-powered significant result may not amount to any evidence for H1 over H0, matching the Bayesian conclusion. Finally the fifth study illustrates that different theories can be evidentially supported to different degrees by the same data, a fact that p-values cannot reflect but Bayes factors can. It is argued that appropriate conclusions match the Bayesian inferences, but not those based on significance testing, where they disagree.

AB - Inference using significance testing and Bayes factors is compared and contrasted in five case studies based on real research. The first study illustrates that the methods will often agree, both in motivating researchers to conclude that H1 is supported better than H0 and the other way round, that H0 is better supported than H1. The next four however, show that the methods will also often disagree. In these cases, the aim of the paper will be to motivate the sensible evidential conclusion and then see which approach matches those intuitions. Specifically, it is shown that a high-powered non-significant result is consistent with no evidence for H0 over H1 worth mentioning, which a Bayes factor can show; and conversely that a low-powered non-significant result is consistent with substantial evidence for H0 over H1, again indicated by Bayesian analyses. The fourth study illustrates that a high-powered significant result may not amount to any evidence for H1 over H0, matching the Bayesian conclusion. Finally the fifth study illustrates that different theories can be evidentially supported to different degrees by the same data, a fact that p-values cannot reflect but Bayes factors can. It is argued that appropriate conclusions match the Bayesian inferences, but not those based on significance testing, where they disagree.

KW - Bayesian analysis

KW - Bayes factors

KW - Replication

KW - Bayes factor

KW - Bayesian statistics

KW - Power

KW - Significance testing

KW - Statistics

U2 - 10.3758/s13423-017-1266-z

DO - 10.3758/s13423-017-1266-z

M3 - Journal article

VL - 25

SP - 207

EP - 218

JO - Psychonomic Bulletin and Review

JF - Psychonomic Bulletin and Review

SN - 1069-9384

IS - 1

ER -