Why repeated measures output gives different output than MANOVA (SPSS)?

Question

I see that this topic was closed few days ago, because it concerns specific context (data). No actually it does not. This is something that happens with any two variables that u have (at least 2) and want do manova on them.

If u try repeated measures design, it will give two parts of analysis: manova, and within subjects

the manova part of repeated measures output is not identical to that when run manova alone. Shouldn't they be the same?

To make it clear: when u have two variables: V1 measured at time1, and V2 measured at time 2 .. and suppose that they r measured on both males and females.

Now u may try RM design. It will show u Multivariate Results and Within Subjects Results.

Or u can try MANOVA analysis directly ...

My issue is. Why the result of the multivariate output in RM analysis isn't identical to that when run MANOVA alone?

Answer 1

Yes I think I found the answer myself. I was deeply thinking.

With RM design, the multivariate analysis splits SS into two components: between subjects (which is usually unwanted) + within subjects So the SS equation involves two components

If we run MANOVA analysis, there is no such split for the SS. It's just between subjects. And therefore the SS has one component only, so naturally, a bit different results of p-values might happen