I am currently running a doubly repeated measures MANOVA . That is, I have two dependent variables being measured over three time points. This produces two multivariate tables. I was hoping someone could explain to me the differences between these two. Also, if I were to pretend instead that these six data points are in fact six dependent variables measured on the same scale simultaneously and run a standard MANOVA, the multivariate table output will be different than the two found in the repeated measures MANOVA. I understand that a MANOVA is incorporating correlation among the dependent variables and therefore uses matrices to establish eigenvalues that are converted to f-values. Repeated measures ANOVA f-values are conceptually derived from a ratio of mean square model to mean square residual (error). So I understand why the f-values found in the univariate output tables are different from those in the multivariate. I am having difficulty finding resources that provide detailed meaning of all the SPSS tables related to my question. Again, I would greatly appreciate any resources that will help me interpret all three of these multivariate tables.

Hi Support Member​ 

The ANOVA and MANOVA commands still exist within the current releases of IBM SPSS Statistics, but now they are aliased to the GLM procedure. Perhaps the references you seeks will be found with searches using "GLM" rather than ANOVA or MANOVA.

Here is the link to the IBM SPSS Statistics Algorithms Guide:

ftp://public.dhe.ibm.com/software/analytics/spss/documentation/statistics/27.0/en/client/Manuals/IBM_SPSS_Statistics_Algorithms.pdf

Discussion of the GLM procedure begins on page 556. Cited references are at the end of each chapter (page 587, for GLM)

I hope this helps!

Support Member​