A test used after an ANOVA to determine which sample means differ significantly from one other, which allows you to protect the results from Type I errors for a small number of comparisons.