2
$\begingroup$

In the dataset I have, there's one categorical variable (A) containing multiple levels. Due to multiple reasons, I want to use continuous variables only for modeling. I feel two other continuous variables (B, C) combined should be good proxies of A. In other words, I want to prove B and C combined are significantly different among levels in A.

So far, I have used MANOVA and got pretty good results. However, I feel MANOVA is more designed for testing if A has an effect on B and C, and here I don't have any causality involved. Is MANOVA the way to go?

$\endgroup$
1
  • 1
    $\begingroup$ Nothing about MANOVA is inherently causal. $\endgroup$ Commented Nov 13, 2014 at 0:47

0

Your Answer

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

Browse other questions tagged or ask your own question.