Dear all,

I am trying to perform a MANOVA with Statistica.  I have got 3 categorical independent variables (named: A, B and C) and 5 dependent numeric variables (named 1,2,3,4 and 5).  The MANOVA, I use GLM module, is significant (Wilks Test has a p <0.00001), but I do not know what I have to do to compare the 3 categorical independent variables. I want to know where and which of the 3 variables is different with the multivariate analysis (Are A, B and C different among themselves? Is only A different of B and C?).

 I also would like to know which of the dependent variables are responsible of the multivariate difference, I would like to underline the dependent variables (if there is any) responsible of the differences, I am looking for  a variable that maximizes the differences among the 3 independent variables using the 5 dependent variables canonical variate. I am looking for standardized regression weights (or standardized discriminant functions coefficients) of the dependent variables responsible of the multivariate difference of the categorical independent variables.

It seems that with the MANOVA of Statistica you only get the Wilks o similar tests significances and then you can only use post hoc or contrast to perform the univariate test or ANOVAS.  I do not know if I have to use the syntax editor to find those analyses for the MANOVA.

I would appreciate your help very much!

Sincerely,

Fco.