MANOVA output- What does this mean?

[Lolly]

As I forgot to ask my supervisor yesterday.

I have run a MANOVA on two DVs, and in the Multivariate Tests table it says that the main effects of Group and Time are not significant, but that the interaction effect is (Group x Time).

I conducted a pre-post intervention study, so I do want a significant interaction effect to show that there was a change in the groups over time.

But I am confused whether I need also to have the main effects of either Time or Group (or both) to be significant.

Does anyone know??

Thanks

 

[Lolly]

Hi,

Can anyone help me on this?

The main effects of Time and Group in my multivariate output are non significant, but the interaction of Time x Group Is significant.

Then when I look at the univariate output, the Time x Group interactions become non significant for the two dependent variables that were entered into the original MANOVA.

I really dont know how to interpret this. What do you make of it?x

 

[Karabiner]

There's no problem with "significant" interactions while main effects are "not signficant".
Regarding the 2nd issue, personally I don't know whether the MANOVA has more power
than a bunch of (uncorrected) univariate ANOVAs; but it seems to me that MANOVA identified
an effect on the multivariate construct (which was built from the single variables), while one
just cannot safely say which of the separate single variables was/were affected...

With kind regards

K.

 

[Lolly]

Thank you Karabiner, you are very kind