Say I have data about a population and I want to analyze them, then in my understanding the "sample" is the population. And then do I need probability, like p-values, confident intervalls etc.?
Given an example: I want to analyse some phenomena in countries and because there are so little (about 200) it is often possible to anaylse all of them. Even if not, I think they are so different that almost every sample would be so biased that make conclusion about all is not approriate (when the sample is random, but that is never the case when analysing countires as far as I know).