Repeated measures ANOVA vs ANCOVA

Hello everyone
Was just wondering, would an output of GLM repeated measures without a covariant (only a factor with 4 levels) be the same for the same factor, regardless of the covariant, when getting a repeated measures ANCOVA output with the same data?

I thought that the only difference whould be that of the covariant row, but it looks like the whole output has been changed, with significant change turning insignificant.



Active Member
The original idea of including the covariate (ANCOVA) was to guard against one of the groups looking different from the others simply because of some external reason that is allowed for by including that covariate. So, one variety of wheat in an experiment might look better than the others until you allow for the natural fertility of the plots. If you include the fertility as a covariate, the differences may easily disappear.
Last edited:
Thanks a lot!
Shouldn't the the interaction (Covariant) × (Factor) as presented in the respective ANCOVA spss output row be significant in such case?
I understand that when the covariate was includud, neither the main effect nor the interaction covariant x main affect are significant? isn't this fact a bit weird, when compared to the No covariant output??


Last edited:


Active Member
I'm a bit confused with the jpg's but as far as the interaction is concerned, a significant interaction in the wheat/fertility example would mean that the effect of fertility was different for each variety. "Not significant" shows that fertility may well influence the yield but it affects each variety in the same way.


TS Contributor
Since you snipped away so much from your output, it is now not very informative, I'm afraid.
I would have liked to see the rows with the error term. Did the covariate have many missing values?

With kind regards