0
$\begingroup$

I am planning an experiment in which I will have a group of animals which I will give both the control diet also test them on the treatment diet.

I think I can use repeated measures ANOVA for this. I'm trying to work out the minimum sample size using G power.

To also look at the effect of the order of diet (whether having the control or treatment diet first or second has any effect). I was planning to split the group in half to two smaller groups and start one group on the control diet and one group on the treatment. Then switch the diets halfway through the experiment (so that both groups have both diets but in opposite orders). ('crossover' design).

I will sample each individual (measure body temperature remotely) once every day for two weeks on each diet to see whether it is affected by diet.

What is the "Number of Measurements"? Would this be the number of times I measure each individual, i.e. 14 or 28? Or is it only 2?

I have found one paper where if I used their method it seems like using 14 or 28 would be correct (https://newprairiepress.org/cgi/viewcontent.cgi?article=1309&context=agstatconference), however in most other examples '2' was used.

If I still only put '2' as the number of measurements then is there any point to measuring the temperature of the animals more than once during each treatment? I feel that measuring the temperature 14 times and finding a difference on most days would be more reliable than if I only took one measurement per animal so I think this should slightly reduce the total sample number required? Hence why I'm wondering if the number of measurements here should be 14.

Is the number of groups 2 because I am using the crossover design?

Any help is greatly appreciated!!

Improve this question
$\endgroup$
4
  • $\begingroup$ Do you assume (presumably on excellent prior evidence) that there is no long term effect of the 'treatment' diet after it is stopped? $\endgroup$
    – BruceET
    Commented Aug 14, 2021 at 17:28
  • $\begingroup$ @BruceET yes that's correct, results of testing this in multiple other studies show it only takes about a week to adjust completely to the new diet $\endgroup$
    – Shara
    Commented Aug 15, 2021 at 2:56
  • $\begingroup$ Of course I guess there could always be an effect that no-one's noticed so that's why I thought it would be good to put half in each group. Also I am taking wild animals into captivity so it's possible that they will be calmer and have lower temperatures after the first month and this will be completely unrelated to diet. Time in captivity was really the main reason I thought it would be good to try to reduce/test for any effect by starting half on the control and half on the treatment. (We will release them again after two months). $\endgroup$
    – Shara
    Commented Aug 15, 2021 at 3:06
  • $\begingroup$ .... p.s. neither the treatment nor control are particularly hard on the animals, just change in fatty acid ratios. $\endgroup$
    – Shara
    Commented Aug 15, 2021 at 3:08

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.