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

Background

edit

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

Statistical hypothesis testing is based on rejecting the null hypothesis when the likelihood of the observed data would be low if the null hypothesis were true. 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  , where   is the desired overall alpha level and   is the number of hypotheses.[4] For example, if a trial is testing   hypotheses with a desired overall  , then the Bonferroni correction would test each individual hypothesis at  .

The Bonferroni correction can also be applied as a p-value adjustment: Using that approach, instead of adjusting the alpha level, each p-value is multiplied by the number of tests (with adjusted p-values that exceed 1 then being reduced to 1), and the alpha level is left unchanged. The significance decisions using this approach will be the same as when using the alpha-level adjustment approach.

Definition

edit

Let   be a family of null hypotheses and let   be their corresponding p-values. Let   be the total number of null hypotheses, and let   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  , that is, of making at least one type I error. The Bonferroni correction rejects the null hypothesis for each  , thereby controlling the FWER at  . Proof of this control follows from Boole's inequality, as follows:

 

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

edit

Generalization

edit

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

Confidence intervals

edit

The procedure proposed by Dunn[2] can be used to adjust confidence intervals. If one establishes   confidence intervals, and wishes to have an overall confidence level of  , each individual confidence interval can be adjusted to the level of  .[2]

Continuous problems

edit

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,  , to the prior-to-posterior volume ratio.[7]

Alternatives

edit

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. But 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

edit

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]

Multiple-testing corrections, including the Bonferroni procedure, increase the probability of Type II errors when null hypotheses are false, i.e., they reduce statistical power.[10][9]

References

edit
  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. ^ a b c Dunn, Olive Jean (1961). "Multiple Comparisons Among Means" (PDF). Journal of the American Statistical Association. 56 (293): 52–64. CiteSeerX 10.1.1.309.1277. doi:10.1080/01621459.1961.10482090.
  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.
  4. ^ Miller, Rupert G. (1966). Simultaneous Statistical Inference. Springer. ISBN 9781461381228.
  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. S2CID 22086583.
  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. arXiv:2007.13821. doi:10.1088/1475-7516/2020/10/009. S2CID 220830693.
  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. ^ a b 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.
edit