Specific Comparisons among Means: Overview (3 of 3)

---

Next section: Computing comparisons

Testing Multiple Comparisons
Even if the comparisons are planned in advance, testing more than one comparison results in an increase in the EER. One way to control the EER is to lower the significance level for the PCER. Naturally, the more comparisons that are conducted, the more the PCER has to be lowered. A second way to control the EER is to use Scheffé's test. Scheffé's test controls the EER no matter how many comparisons are conducted.

Next section: Computing comparisons