How am I going to compute for sample Size?

Question

We already had decided on a few importnt things for our research.We are going to use rabbits as our test animal. We are going to compare two groups using a paired t-test.
Each group is composed of 8 rabbits. Here is the flow of our procedure...acclimatize rabbits for 3 days, then get baseline cholesterol level of both groups.next, we are going to fed the 2 groups with pellets coated with lard inorder to increase their serum cholesterol level.Then get the cholesterol level(pre-test). Then for four weeks, we will treat the experimental group with geen tea while the control group for simvastatin which is a cholesterol lowering drug and then compare the result(post-test).

How are we going to compute for the sample size?

How will it differ if were are going to do these just by using 3 rabbits for each group?Or is that possible?

Answer 1

It all depends on how much different you expect the results to be between the experimental group and the control group. I am on simvastatin and have gone periods drinking more green tea. Have not watched my HDL levels closely enough to compared effects.

I use statcalc in EpiInfo to run sample size calculations and find that is easy and accurate. I think that would work for your experiment and your question.

Answer 2

Your control group should be rabbits still eating the lard coated pellets. How are you going to tell if the cholesterol decreased because of the drug, and not because you stopped feeding them lard coated pellets. If you ARE going to keep feeding both groups lard coated pellets, then your tests sample sizes should depend on the P values. The smaller the deviation, the smaller the sample size needed to state with confdence and accuracy