Multiple comparisons



Is there a rule of thumb to help deterimine when multiple comparisons are an issue within a single study/paper/report?

For example, if I use a 2-way ANOVA with two factors on a single variable I can get a number of relevant comparisons using post-hoc tests. However, if I have more than one variable, the ANOVA would be run more than once. Is that an issue?


Super Moderator
If the y's are related it will be and you should be looking at a MANOVA model, if they are completely independent, it should be fine.