0
$\begingroup$

I'm conducting a repeated measures MANOVA with one IV (4 conditions) and 2 DV's. My participants have taken each condition.

  1. Assuming the MANOVA and univariate tests are significant, I would like to conduct planned comparisons instead of post hoc tests. I can only find information about this for ANOVA's and nothing for MANOVA's. Can I do this or do I need to use posthoc tests?

  2. I've also found lots of references to repeated measures MANOVA as a profile analysis. They're not called the same thing but seem to be same. Do they mean the same thing? Or are they different tests?

  3. I would also like to potentially introduce a co-variate (age) and do a repeated measures MANCOA. Again, I can't find any information on this. Could someone point me to an SPSS book with instructions on this?

Thanks so much!

Improve this question
$\endgroup$

1 Answer 1

Reset to default
0
$\begingroup$

I don't really understand why you have these questions. Manova and Mancova are pretty well documented.

Can you explain how this blog:

http://www.statsmakemecry.com/smmctheblog/how-to-conduct-a-repeated-measures-mancova-in-spss.html

is not enough explanation for you for example so we understand better what you need?

Edit after your comment

Related to Q1,

With Manova it is the same as with ANOVA. You hypothesize that there is a difference between groups on some numerical variable influenced by your independent variables/groups.

Why do we first do a MANOVA and then afterwards use post-hoc t-tests to see specific differences?

Because if we do the t-tests before, we are testing multiple hypotheses which causes loss of power, see bonferonni holm method. With only one variable with 5 groups this would already lead to (5x5 -5) : 2 = 10 tests instead of only the one test when doing manova.

If you want to do planned comparisons and are not interested in the full scale comparisons you can do some hypotheses about targeted groups and leave the groups you are not interested in out of your manova. Just keep in mind that with each test you lose power.

Related to the book question:

[John_Fox]_Applied_Regression_Analysis_and_General_liniar_models might help you since manova t-test ancova etc. is all glm. There are also references in the end to articles that might help you out.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thanks. I've found that blog and it is very helpful but I was looking for a stats book that I could cite and with more information. Although stats books do certianly cover MANOVA and MANCOVA they don't usually cover repeated measures MANCOVA. Or at least I haven't found any books that do so far. Any answers to Q1 would also be really helpful. The pairwise tests are posthoc but I would like to run planned comparisons if that's possible. Thanks! $\endgroup$
    – statshelpneeded
    Commented Mar 29, 2020 at 15:14
  • $\begingroup$ I edited my answer $\endgroup$
    – B. Bram
    Commented Mar 31, 2020 at 11:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.