The model of a randomly selected sample from a finite population has been very fruitful. Sophisticated sampling techniques that are both economical and scientifically reliable have been developed. An entire industry of public opinion polling as well as the technical activities of the [U.S. Bureau of the Census] depend on these techniques.
The most elementary methodology is called simple random sampling. It assures that every possible subset of the population which has the desired sample size is given the same probability of selection. The possibility of very expensive or very atypical samples has lead to a variety of modfications such as stratified samples, cluster sampling, and multistage samples. The most experienced center in these techniques outside the Census Bureau is the [University of Michigan Survey Research Center].