I've tried to find an answer to my question by searching the net but I can't find any clear solution.
I want to compare the effect of treatment A with treatment B on the % of 4 different dependent variables within a fixed/closed compartment. I have decided that MANOVA would not be appropriate as the DVs are not entirely separate (i.e. as the % of one goes up, the % of the others will change accordingly) and so decided to perform a simple t test between A and B for each DV.
In total I have performed 4 different t tests. Do I need to correct for the family-wise error rate even though I have not performed an omnibus test (like ANOVA/MANOVA)?
One other point, my hypotheses are whether treatment has an effect on each DV separately.
I hope that makes sense.
Thank you for any replies.
Last edited by hrj21; 06-27-2012 at 02:46 AM. Reason: sinple - simple
Advertise on Talk Stats