I'm aware that a t-test could be used between means at each time-point; but isn't it wrong to analyse each time point separately (increase type 1 error?).
Do I have to include time in as a factor, in which case do I have to use an ANCOVA? Or a factorial ANCOVCA? I'm also reading of linear mixed effect modelling....
Any help figuring out the right test will be really appreciated! Many thanks!
A Google search on should bring up a page of examples using your software (assuming it is one of the major packages). Look especially at the examples that use a multilevel modeling approach. HTH.
I would recommend Prof Marie Davidian's course notes on longitudinal statistical models. These are available on her website through the Stat department of NCSU. GOOGLE all of this to find their current URL
Best, David Booth
Hello, I think you can manage to solve your problem by using Repeated Measures ANOVA with an within (time) and a between factor (groups). If you have a lot of time points (for example 10+) it would be better to use a longitudinal approach.
I advise you tu use Greenhouse-Geisser or Huynd-Feldt correction in case of sphericity violation.
i recently put together a little note regarding a similar situation (pre-post only, but the extension to multiple time-points is straightforward). maybe it helps:
Method Model representations and sample size for a two arm pre-post...
As you have multiple measurement times on the same two groups, you should use repeated measure ANOVA here. See here how to perform repeated measure ANOVA
https://www.reneshbedre.com/blog/repeated-measure-anova.html
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size