ANOVA follow-up tests

Hi everybody,

This is more of a theoretical question when using the statistical toolbox:

1. Is it correct to that you're not allowed to do follow-up tests on the your ANOVA unless you see an interaction?

2. Is it OK to use t-tests as follow-up tests? Or are there better solutions in matlab?

Best regards,

Christian

Answer 1

[Star Strider]

I doubt the ANOVA Police will come after you if you don’t see an interaction. They may wonder why you’re doing multiple comparisons if you don’t have a statistically-significant result.

Doing t-tests with the multcompare function is a common practice.

[Eric]

I just don't want to be in bad standing with the ANOVA police since I might have to violate the rules later. :)

Thanks!

[Star Strider]

My pleasure!