Combination of Probabilities from Significance Tests

Abstract

PROBABILITIES from independent significance tests, in which usually none of the separate probabilities is significant, are usually pooled by Fisher's method1, which is based on the product of the individual probabilities. Although Fisher states that a χ2 transformation provides an exact test of significance in this situation, Wallis2 pointed out some time ago that this is only true if the separate probabilities are derived from continuous distributions. When the original distributions comprise a limited number of finite probabilities, Fisher's transformation often results in a considerable over-evaluation of the true probability of the product.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

References

  1. 1

    Fisher, R. A., Statistical Methods for Research Workers, thirteenth ed. revised (Oliver and Boyd, Edinburgh, 1958).

  2. 2

    Wallis, W. A., Econometrica, 10, 229 (1942).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

MCLAUGHLIN, R. Combination of Probabilities from Significance Tests. Nature 220, 1250–1251 (1968). https://doi.org/10.1038/2201250a0

Download citation

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.