maov2

Computes a two-way Multivariate Analysis of Variance for equal or unequal sample sizes.

Computing of a two-way multivariate analysis of variance for equal or unequal sample sizes by the testing of the mean differences in several variables among several samples with two factors.

It consider the matrices of sum-of-squares between (H) and within (E) for main effects and their interaction to compute, respectively, their Wilks'L (lambda) and according to the number of data it chooses to approximate to the Chi-square distribution (large samples: n >= 25) or through several multiple F-approximations such as Rao, Pillai, Lawley-Hotelling and Roy tests (small samples: n < 25).

A warning exists if likelihood ratio procedures are not satisfied. Interaction is carried out before the tests for main effects. If interaction effects exist, the factor effects do not have a clear interpretation and the file ask you whether it proceed with the additional multivariate tests.

It needs to input data matrix (size of matrix must be n-by-(2+p); factor 1=column 1; factor 2=column 2, variables=column 3:p) and significance (default = 0.05).

The output is a complete multivariate analysis of variance table for each factor and their interaction.