anova

JohnM

TS Contributor
#2
You would either do multiple comparisons after an overall F-test with ANOVA, or do them at the start without ANOVA, if your hypothesis specifically calls for it.

They may or may not be significant, depending on the effect size, among other factors.

The Dunnet test is a special type of comparison (a control group vs others), and can't really be directly compared against Tukey and Newman-Keuls.

Please check the Online Resources links in the Examples forum, and they should lead you to some information on multiple comparison tests.