Is there anything that would prevent me from using the same test firstly for stratifying the subjects then secondly for measuring the change after administering the treatment?
It depends on why you are stratifying your sample. What will happen to your three groups after you stratify them?
I will stratify them into three categories according to a specific functional impairment they have. Under each category, the subjects will be randomly assigned to either experimental or control group.
Thank you very much for your reply
As long as you are examining each group separately, there should be no problem with using one of your tests as a pre- and post-assessment. Each group's analysis is a 2 group repeated measures (2x2 structure).
If you want to look across the full sample, the model is more complicated, since you have 3 levels, 2 groups, and 2 time-points. But it should still be do-able. The problem you have is that the growth on the measure is not independent of the level membership, so you may not find effects (Type 2 error). I would look at formal growth curve modeling for a full sample analysis.
- Badges
- Science topic
- Similar topics
- Computer Science
- Research Topics