I got a stupid isue and not sure how to interpret it.
I conducted a MANOVA (1 categorical [Range 1-5] IV; 2 intervall DVs] and got a significant Wilks' Lambda (same for the other 3 multivariate tests). Wilks' Lambda is quite high though with .834.
My two DVs show positive linear correlation with Pearson's correlation coefficient being r(89)=.31; p=.003. So perfectly suitable for a MANOVA one should think.
Then I did what 90% of all people do (despite the criticism associated with it) and had two follow up ANOVAs run over the data. Interestingly none of them indicated significant results.
From what I've gathered this seems to be a rare occurence and I have no idea what to make of it. Where comes the significance in my MANOVA from? And why doesn't it show in the ANOVAs? Given how using ANOVAs as follow up methods for MANOVAs is considered to increase the chance of a type I error I expected quite different results.
So how do I interpret that? A significant MANOVa yet no significant ANOVAs.
I am aware that I could use a DFA/DA or its equivalents to get a better idea of the data yet inverting IV and DVs doesnät really make sense in my design and even when doing it for the sake of being able to analyse the data further I don't see how it helps me to explain why I got a significant MANOVA but no significant ANOVAs.
If anyone could help, that' be great!