Hi, I'm going to use pre- and post-test ANCOVA to analyse the results of an intervention study. I have some time specific variables, such as number of the children present in the post-test video observation, that I would like to use as a covariate in SPSS analyses. Is this allowed considering that the number of the children varies between the measurement points? The pretest scores of the dependent variable is also used as a covariate.
Hello Sivi,
In an ordinary ancova, the pre-treatment measure is the one that should be used as the covariate. So for your example, number of children present at baseline. Since it is possible that the treatment, other variables, or some interaction of same could impact the post-intervention count (of children), that number should not be used as a covariate.
The exception to this would be if you were trying to develop a growth model that had multiple observation times beyond two. In such a model, values at time 2 ("immediate?" post treatment) could be used as a covariate/control variable for values at time 3, and so on.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS