The Bonferroni Method for Multiple Comparisons

When the differences of average levels among several groups are statistically significant, the multiple comparisons should be carried out. Using α=0.05 directly as significance level in comparison between each pair may result in too high typeⅠ error probability. Bonferroni method is one of the most common methods for multiple comparisons, which could reduce the accumulated typeⅠ error probability by adjusting α level. In this report, the fundamental principle of Bonferroni method is explained and examples are presented in some specific situations. This article is supposed to guide researchers to utilize Bonferroni method correctly in multiple comparisons.