Chapter 4 Sec 1.3: Sampling from a Finite Population
So far, we have treated each set of data that we have investigated as a sample from an ongoing random process with a constant (but unknown) underlying probability of success. It is more common to consider sample data as coming from a larger, but finite population (e.g., a sample of adults from among all Americans aged 18 years and older). So what we now want to consider is (i) how to select a sample from the population to allow us to generalize our sample observations back to that larger population and (ii) what statistical techniques we need to make such inferences. You will see that though many of the analysis procedures are the same, there are a few more issues to consider as well.
