Benferroni Correction

Jump to: navigation, search

See: Benferroni, Correction, Statistical Test.



    • BONFERRONI CORRECTION When n statistical tests are carried out, each has the potential (probability, p, the significance level) to return a false-positive result. If tests are independent of each other, the so-called experiment-wise probability that one or more tests show a false-positive result is approximately np. So, to achieve an experiment-wise false-positive rate of p, each individual test must only be allowed a false-positive error rate of p/n, which is referred to as the Bonferroni correction.