3 Dependent and 4 Independent Variables...MANOVA?

Hello all,

My study recorded 3 continuous dependent variables in relation to an ordinal experimental factor. However, there are 3 categorical bias factors that may have effected the results (some preliminary analysis leads me to believe that only 1 of these had any significant effect).

I'm unsure if I can run a MANOVA on this data as I can only find references to up to two-way MANOVAs and this would be a 4-way. Can one perform 4-way MANOVAs? (analyze ->general linear model ->multivariate?)

Also, is there an alternative to MANOVA that does not assume equality of variances?



I also considered running forward model selection to find the best model and then just running a two-way/one-way MANOVA based on that (as I said I expect 2 of the bias factors to drop out). Is that an okay methodology?
Last edited: