Hi, I'm doing research on farmed pigs. The aim is to observe the animals and classify each one as poor/good welfare at 5 different points in their lives. The aim is to determine whether animals showing good/poor welfare in early life, later life, all through life have specific carcass traits when they are slaughtered. The aim is not to accurately reflect the prevalence of the various health issues within the entire population. So I'm wondering in a case like this where inferring back to the population is not the goal, can a larger margin of error (say 10) be used for estimating sample size?
Thank you!
Thank you!