What test to use?

#1
Hi,

I have 2 groups with n=6 each with respiratory rate taken at baseline (predose), 1, 2, 3, and 4 hour postdose. I have figures with baseline subtracted means and SD at each timepoint and need to know whether there is a difference between groups. Can it be as simple as running multiple t tests at each timepoint using the baseline subtracted values or would it be best to do a two way anova and compare against the baseline values for each group?
 

hlsmith

Not a robit
#4
If they are randomized do you need to subtracted out baseline RR, shouldn't these values be comparable given the design?
 
Last edited:
#6
Maybe I can take the baseline subtracted values for both groups and run two way rm anova (time and treatment) to compare the two groups at each timepoint...would that make sense?
 

hlsmith

Not a robit
#7
The sample size is just so small, you are going to need a big difference especially after controlling for multiple tests (false discovery). You will also need to test assumptions for the model (e.g., linearity and residuals). I want to say conduct a corrected permutation test, but given your sample size I not even sure that may work.

P.S., It is traditional to control for the baseline value in the model, not subtract it out - but your model is lacking degrees of freedom.