MANOVA with 2 repeated measures (Effect of Time and Group)

[Dusan Petrovic]

Dear statalisters,

I am performing manova analysis with repeated measures (2 time points)

so the model I am implementing is the following :

qui manova time1 time2 = i.Groupe2 i.sex c.age

//I am only intrested in the effects of time and Groupe2 (3 levels) , but I'm adjusting for sex and age
// on this page, I found exactly what I want : http://www.stata.com/manuals13/mvmanova.pdf (pp. 23 and after)
// So I'm proceeding like this :

matrix m1=J(1,2,1)                // (2 : indicates number of time levels)
matrix m2=(1,-1)                   // Comparison time points (time point 1 vs 2) : 1 row => 1 comparison


 
 matrix c1=(1,-1,0,0,0,0 \ 0,1,-1,0,0,0 \ 1,0,-1,0,0,0) // Comparison Groupe2 
 // 3 rows => 3 comparaisons.
//  6 cols    (3 levels Groupe2 + 2 levels Time + 1 cons)
 
 matrix c2=(0.333,0.333,0.333,0.5,0.5,1,1) // Weight for independent variables (Groupe2, sex01, AGE, CONS)

  
 qui manovatest, test(c1_`i') ytransform(m1_`i') // EFFECT GROUPE
 *matrix gr=r(stat)
 
 qui manovatest, test(c2_`i') ytransform(m2_`i') // EFFECT TIME
 *matrix time=r(stat)

 
 qui manovatest, test(c1_`i') ytransform(m2_`i') // INTERACTION GROUPE-TIME
*matrix grtime=r(stat)

Unfortunately, there is a conformability error.
I personnally think that it is either coming from m2 or c1 matrices, but I cannot figure it out...
I also have a model with 3 time point for which it works.



Your kind help is really appreciated

[Dusan Petrovic]

Don't worry about the "_`i' " it's because I was using this in a loop