The Tukey test uses the Studentized range statistic to make all pairwise comparisons between groups. The test is based on the Tukey-Kramer method and sets the experimentwise error rate at the error rate for the collection for all pairwise comparisons. The Tukey test is most effective when testing a large number of pairs of means.

The test performs well by keeping the significance level of the simultaneous pairwise tests at the prescribed level. For homogeneous subset testing, the test uses the harmonic sample mean Qh and exhibits large errors in reported significance level for strongly unbalanced designs where the group sample sizes are very different. The test is not recommended for use with such designs.