126 PART TwO • THE POlITICS OF AMERICAn dEMOCRACy
Several newcomers to the public opinion poll industry accurately predicted Roosevelt’s
landslide victory. These newcomers are still active in the poll-taking industry today: the
Gallup poll of George Gallup and the Roper poll, founded by Elmo Roper. Gallup and
Roper, along with Archibald Crossley, developed the modern polling techniques of market
research. Using personal interviews with small samples of selected voters (fewer than two
thousand), they showed that they could predict with relative accuracy the behavior of the
total voting population.Sampling Techniques
How can interviewing fewer than two thousand voters tell us what tens of millions of vot-
ers will do? Clearly, it is necessary that the sample of individuals be representative of all
voters in the population.
The most important principle in sampling, or poll taking, is randomness. Every per-
son should have a known chance, and especially an equal chance, of being sampled. If
sampling follows this principle, then a small sample should be representative of the whole
group, both in demographic characteristics (age, religion, race, region, and the like) and
in opinions. The ideal way to sample the voting population of the United States would
be to put all voter names into a jar—or a computer file—and randomly sample, say, two
thousand of them. Because this is too costly and inefficient, pollsters have developed other
ways to obtain good samples. One technique is simply to choose a random selection of
telephone numbers and interview the respective households. This technique used to pro-
duce a relatively accurate sample at a low cost.
To ensure that the random samples include respondents from relevant segments of
the population—rural, urban, northeastern, southern, and the like—most survey organi-
zations randomly choose, say, urban areas that they will consider as representative of all
urban areas. Then they randomly select their respondents within those areas.The Statistical nature of Polling. Universally, when the results of an opinion poll
are announced, the findings are reported as specific
numbers. A poll might find, for example, that
10 percent of those surveyed approve of the job
performance of Congress. Such precise figures
can mislead you as to the essential nature of
polling. In reality, it makes more sense to con-
sider the results of a particular survey question
as a range of numbers, not a single integer.
That would mean that the question
about Congress’s job performance
yielded an answer that fell some-
where between 7 percent and
13 percent. The figure of 10 per-
cent is only the midpoint of the
possible spread—the most prob-
able result. If we had been able to
question all members of the pub-
lic, the chances that they would
give Congress exactly a 10 per-
cent rating are not high. Even if
the pollster in this case employedHarry Truman won the presidential election in 1948 despite the prediction of
most opinion polls that he would lose. Would a newspaper today make such an inaccurate
prediction and put it on newsstands? (AP Photo/Byron Rollins)Copyright 2014 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.