I originally thought this would be a MANCOVA, but after researching this I can't use categorical covariates.
The assumption to use MANCOVA is that covariance must hold between variables. This is you already arriving at this assumption, then go ahead. Mind you, MANOVA comes with the opposite assumption
Dear Katie Minnett , good question!
The fact that an explanatory variable (i.e., independent variable) is categorical or not does not impact at all the use of MANOVA/MANCOVA.
In a fully-fixed effects models, there is very little benefit in using MANOVA/MANCOVA. However, it can give you the opportunity to test if the effect of the explanatory variable differs between response (i.e., dependent) variables, which seems to be your interest.
Is this the case? What do you mean by "comparing 3 DV's"? Comparing them with regards to what?
Nonetheless, you are good! Thanks, Nick
Hello Katie,
The proposed categorical covariates can be handled either as: (a) a traditional covariate, if dichotomous (and coded as 0/1, or 1/2, for example), or, for three or more levels, as a set of k - 1 dummy variates (where k is number of levels); (b) as blocking factors in the model (if you're considering an m/anova framework).
The multivariate regression analog (using dummy variables) can also be used to answer the same questions.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Cognitive Science
- Thinking