Imagine this experiment: assume that I have two sets of twins. For the first pair of twins, I keep one as a control, and the second twin I give drug A. For the second pair of twins, I keep one as a control and I give the second twin drug B. Now, I want to show that the effectiveness of drug A over control is bigger than the effect of drug B over control. In each case, I measure a response in each twin like say body temperature.
My goal would be to prove that drug A lowered body temperature more than drug B compared to the control, a single p-value summarizing the experiment for instance.
I replicate the whole experiment three times. So, for drug A, I have three pairs of numbers, each making a comparison between a pair of twins, one given the drug and one left as a control; and the same for drug B.
How do I go about analyzing an experiment like this?
(For a lot of technical reasons related to my application, I have to work with matched pairs; and I'm not really experimenting on humans.)