Hi All,
In repeated measure analysis, what matters is whether there is a significant group*time interaction or not. Do you think adjusting for pre-test is required or logical while we are testing if the pattern of change is different for the groups? If so, how can I perform it in repeated measure MANCOVA using SPSS as I have several variables?
Any adjustment to the pretest would introduce your or someone's bias into the process and change the growth patterns, thus changing the interactions. I would not recommend that. There is one exception. If different forms of the pretest instrument were used, the forms could be equated using traditional or IRT equating methods.
An important factor to consider is if the pre-score influenced the allocation into your groups. Can you say more about what your groups are and if this was the case? If it did not using a covariate can be problematic.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS