Bonferroni correction

From HandWiki
Short description: Statistical technique used to correct for multiple comparisons

In statistics, the Bonferroni correction is a method to counteract the multiple comparisons problem.

Background

The method is named for its use of the Bonferroni inequalities.[1] An extension of the method to confidence intervals was proposed by Olive Jean Dunn.[2]

Statistical hypothesis testing is based on rejecting the null hypothesis if the likelihood of the observed data under the null hypotheses is low. If multiple hypotheses are tested, the probability of observing a rare event increases, and therefore, the likelihood of incorrectly rejecting a null hypothesis (i.e., making a Type I error) increases.[3]

The Bonferroni correction compensates for that increase by testing each individual hypothesis at a significance level of [math]\displaystyle{ \alpha/m }[/math], where [math]\displaystyle{ \alpha }[/math] is the desired overall alpha level and [math]\displaystyle{ m }[/math] is the number of hypotheses.[4] For example, if a trial is testing [math]\displaystyle{ m = 20 }[/math] hypotheses with a desired [math]\displaystyle{ \alpha = 0.05 }[/math], then the Bonferroni correction would test each individual hypothesis at [math]\displaystyle{ \alpha = 0.05/20 = 0.0025 }[/math]. Likewise, when constructing multiple confidence intervals the same phenomenon appears.

Definition

Let [math]\displaystyle{ H_1,\ldots,H_m }[/math] be a family of null hypotheses and [math]\displaystyle{ p_1,\ldots,p_m }[/math] their corresponding p-values. Let [math]\displaystyle{ m }[/math] be the total number of null hypotheses, and let [math]\displaystyle{ m_0 }[/math] be the number of true null hypotheses (which is presumably unknown to the researcher). The family-wise error rate (FWER) is the probability of rejecting at least one true [math]\displaystyle{ H_{i} }[/math], that is, of making at least one type I error. The Bonferroni correction rejects the null hypothesis for each [math]\displaystyle{ p_i\leq\frac \alpha m }[/math], thereby controlling the FWER at [math]\displaystyle{ \leq \alpha }[/math]. Proof of this control follows from Boole's inequality, as follows:

[math]\displaystyle{ \text{FWER} = P\left\{ \bigcup_{i=1}^{m_0}\left(p_i\leq\frac \alpha m \right) \right\} \leq\sum_{i=1}^{m_0}\left\{P\left(p_i\leq\frac \alpha m\right)\right\} = m_0 \frac \alpha m \leq \alpha. }[/math]

This control does not require any assumptions about dependence among the p-values or about how many of the null hypotheses are true.[5]

Extensions

Generalization

Rather than testing each hypothesis at the [math]\displaystyle{ \alpha/m }[/math] level, the hypotheses may be tested at any other combination of levels that add up to [math]\displaystyle{ \alpha }[/math], provided that the level of each test is decided before looking at the data.[6] For example, for two hypothesis tests, an overall [math]\displaystyle{ \alpha }[/math] of 0.05 could be maintained by conducting one test at 0.04 and the other at 0.01.

Confidence intervals

The procedure proposed by Dunn[2] can be used to adjust confidence intervals. If one establishes [math]\displaystyle{ m }[/math] confidence intervals, and wishes to have an overall confidence level of [math]\displaystyle{ 1-\alpha }[/math], each individual confidence interval can be adjusted to the level of [math]\displaystyle{ 1-\frac{\alpha}{m} }[/math].[2]

Continuous problems

When searching for a signal in a continuous parameter space there can also be a problem of multiple comparisons, or look-elsewhere effect. For example, a physicist might be looking to discover a particle of unknown mass by considering a large range of masses; this was the case during the Nobel Prize winning detection of the Higgs boson. In such cases, one can apply a continuous generalization of the Bonferroni correction by employing Bayesian logic to relate the effective number of trials, [math]\displaystyle{ m }[/math], to the prior-to-posterior volume ratio.[7]

Alternatives

There are alternative ways to control the family-wise error rate. For example, the Holm–Bonferroni method and the Šidák correction are universally more powerful procedures than the Bonferroni correction, meaning that they are always at least as powerful. Unlike the Bonferroni procedure, these methods do not control the expected number of Type I errors per family (the per-family Type I error rate).[8]

Criticism

With respect to FWER control, the Bonferroni correction can be conservative if there are a large number of tests and/or the test statistics are positively correlated.[9]

The correction comes at the cost of increasing the probability of producing false negatives, i.e., reducing statistical power.[10][9] There is not a definitive consensus on how to define a family in all cases, and adjusted test results may vary depending on the number of tests included in the family of hypotheses.[citation needed] Such criticisms apply to FWER control in general, and are not specific to the Bonferroni correction.

References

  1. ↑ Bonferroni, C. E., Teoria statistica delle classi e calcolo delle probabilità, Pubblicazioni del R Istituto Superiore di Scienze Economiche e Commerciali di Firenze 1936
  2. ↑ 2.0 2.1 2.2 Dunn, Olive Jean (1961). "Multiple Comparisons Among Means". Journal of the American Statistical Association 56 (293): 52–64. doi:10.1080/01621459.1961.10482090. http://sci2s.ugr.es/keel/pdf/algorithm/articulo/1961-Bonferroni_Dunn-JASA.pdf. 
  3. ↑ Mittelhammer, Ron C.; Judge, George G.; Miller, Douglas J. (2000). Econometric Foundations. Cambridge University Press. pp. 73–74. ISBN 978-0-521-62394-0. https://books.google.com/books?id=fycmsfkK6RQC&pg=PA73. 
  4. ↑ Miller, Rupert G. (1966). Simultaneous Statistical Inference. Springer. ISBN 9781461381228. https://books.google.com/books?id=4C7VBwAAQBAJ. 
  5. ↑ Goeman, Jelle J.; Solari, Aldo (2014). "Multiple Hypothesis Testing in Genomics". Statistics in Medicine 33 (11): 1946–1978. doi:10.1002/sim.6082. PMID 24399688. 
  6. ↑ Neuwald, AF; Green, P (1994). "Detecting patterns in protein sequences". J. Mol. Biol. 239 (5): 698–712. doi:10.1006/jmbi.1994.1407. PMID 8014990. 
  7. ↑ Bayer, Adrian E.; Seljak, Uroš (2020). "The look-elsewhere effect from a unified Bayesian and frequentist perspective". Journal of Cosmology and Astroparticle Physics 2020 (10): 009. doi:10.1088/1475-7516/2020/10/009. https://doi.org/10.1088%2F1475-7516%2F2020%2F10%2F009. 
  8. ↑ Frane, Andrew (2015). "Are per-family Type I error rates relevant in social and behavioral science?". Journal of Modern Applied Statistical Methods 14 (1): 12–23. doi:10.22237/jmasm/1430453040. 
  9. ↑ 9.0 9.1 Moran, Matthew (2003). "Arguments for rejecting the sequential Bonferroni in ecological studies". Oikos 100 (2): 403–405. doi:10.1034/j.1600-0706.2003.12010.x. 
  10. ↑ Nakagawa, Shinichi (2004). "A farewell to Bonferroni: the problems of low statistical power and publication bias". Behavioral Ecology 15 (6): 1044–1045. doi:10.1093/beheco/arh107. https://academic.oup.com/beheco/article/15/6/1044/206216. 

External links



Retrieved from "https://handwiki.org/wiki/index.php?title=Bonferroni_correction&oldid=3383856"