We are using Giving Up Densities to determine the perceived risk in voles.
This is a forced choice setting, with 3 different foraging patches, each with their own risk level.
Some individuals started to forage in the "dangerous" patches and store it in the "safe" patches, bringing them above their initial food levels.
How do I account for this in the statistics?
We are going to report proportions (food remaining/food initially), so that values should range between 0 and 1, however the latest individual stored so much food that it pushes the value to 1.2.
My idea was to either force a cut-off at 1, or to subtract the additional food, i.e. turn the 1.2 into 0.8.
Neither option seems perfect.
Thanks in advance!
Dear Thorbjörn,
Is the relocation of forage in your experiment common? How long did trails last?
I think ideally, you would not want to measure any relocated food, but only what has been removed from a certain location. Therefore, I would assume that subtracting the additional food is the way to go.
Best regards
Bjorn
Hey Bjorn Mols ,
I can identify 8 samples with relocated food (88 animals, 3 samples each, 264 samples total) from the data. The trials lasted 2h.
I now have all the data and tested both options. As it turns out, the results do not even differ much from each other.
I think I will keep subtracting the additional food to not underestimate the foraging.
This is a really interesting question, statistically. I think it would be wrong to subtract the additional food that the voles moved to the safe patch, since the voles react to the amount of food that's actually in the patch, not the amount you put there. One way around it would be to treat it as a longitudinal experiment, with food levels in the different patches changing through time since start of the experiment (with time as a random factor).
- Badges
- Science topic
- Similar topics
- Medicine
- Nutrition and Dietetics
- Nutritional Medicine
- Food Science
- Food