ABSTRACT
Nothing is more important in surveys than sampling. The catastrophe of the Literary Digest survey that predicted the certain defeat of Franklin D. Roosevelt could be traced to errors in its sampling techniques, a lesson occasionally forgotten in the expediency of events or time. Likewise, the Chicago Tribune's publication of the resounding defeat of Harry S. Truman was another classic example of sampling errors. It can be flatly stated that any survey in voting, advertising, marketing, or communication is only as good as its sample. There is no good survey without a good sample. As the foundation of any survey, sampling determines whether the results are accurate, valid, reliable, and representative. The importance of sampling has been so repeatedly emphasized as to sound like a parrot's talk. Any sample, regardless how it is drawn, contains errors, because it represents the population but is NOT the population. Shall we say: it represents the real thing but is not the real thing!
