However, there is an alternative approach, based on the plausible assumption that the population should look roughly like the sample. Since we cannot repeatedly draw a new sample from the population, we instead repeatedly draw new samples from our sample!
