Box's test in MANOVA

#1
Hi,
I have run MANOVA with 3 groups in the independent variables and 8 variables for the dependent variables.
I have 4, 5, 5 subjects in each group respectively.

In the output, I get this message.

Box's Test of Equality of Covariance Matrices is not computed because there are fewer than two nonsingular cell covariance matrices.

I appreciate if someone can tell me what this means.

Thank you,
 

spunky

Doesn't actually exist
#2
you cannot run a box test when you have less subjects per factor than dependent variables... Box's M test assumes that the covariance matrices are invertible (i.e. nonsingular). in your case i guess most covariance-matrices-per-factor are singular (i.e. they do not have an inverse).
 
#4
My data is almost identical in design and I got the same error message about Box's Test. Is there another way to test for the assumption of equality of covariance matrices?