Discovering Statistics

mixed design MANOVA

Viewing 1 post (of 1 total)
  • Author
    Posts
  • #628

    Dear all,

    I’m trying to perform in R mixed design manova with one between subject variable and one within subject variable.

    I wrote the following lines in order to obtain the results.

    options(contrasts=c("contr.sum","contr.poly"))
    manova.out = manova(cbind(DV1,VD2) ~ IV*groups + Error(id/IV), data=longdata1)
    summary(manova.out, test="Wilks")

    Given that it was the first time I tried this analysis in R, I wanted to compare my results with those obtained in statistica and SPSS. This is where my problems appeared.

    In the attached pdf, I added the results of R, statistica and SPSS.

    Concerning the between subjet factor, the three softwarse obtained the same results. Good news!
    However, it is not the case for the within subjet factor and for the interaction effect, bad news!

    Statistica and the first SPSS table match... but the within multivariate effect (second SPSS table) matches with R results. Thus, I'm confused. Are both SPSS tables presenting the results of the same hypotheses by different algorithm ? Which of the SPSS table is the correct one ? Did I make some mistakes in defining my manova model in R? If so, does anyone can help me to write it correctly ?

    Thanks for your help,
    Nicolas
Viewing 1 post (of 1 total)
  • You must be logged in to reply to this topic.