The Language of Argument

1 8 7

S t a t i s t i c a l G e n e r a l i z a t i o n s

What went wrong this time? The answer here is more subtle. The Gallup

pollsters (and others) went to great pains to make sure that their sample was

representative of the voting population. The interviewers were told to poll a

certain number of people from particular social groups—rural poor, subur-

ban middle class, urban middle class, ethnic minorities, and so on—so that

the proportions of those interviewed matched, as closely as possible, the pro-

portions of those likely to vote. (The Literary Digest went bankrupt after its

incorrect prediction, so the pollsters were taking no chances.) Yet somehow

bias crept into the sampling; the question was, “How?” One speculation was

that a large percentage of those sampled did not tell the truth when they

were interviewed; another was that a large number of people changed their

minds at the last minute. So perhaps the data collected were not reliable. The

explanation generally accepted was more subtle. Although Gallup’s work-

ers were told to interview specific numbers of people from particular classes

(so many from the suburbs, for example), they were not instructed to choose

people randomly from within each group. Without seriously thinking about

it, they tended to go to “nicer” neighborhoods and interview “nicer” people.

Because of this, they biased the sample in the direction of their own (largely)

middle-class preferences and, as a result, under- represented constituencies

that would give Truman his unexpected victory.

Is the Sampling Procedure Biased?

Because professionals using modern techniques can make bad statistical

generalizations through biased sampling, it is not surprising that our every-

day, informal inductive generalizations are often inaccurate. Sometimes we

go astray because of small samples and biased samples. This happens, for

example, when we form opinions about what people think or what people

are like by asking only our friends. But bias can affect our reasoning in other

ways as well.

One of the main sources of bias in everyday life is prejudice. Even if we

sample a wide enough range of cases, we often reinterpret what we hear or

see in light of some preconception. People who are prejudiced will find very

little good and a great deal bad in those they despise, no matter how these

people actually behave. In fact, most people are a mixture of good and bad

qualities. By ignoring the former and dwelling on the latter, it is easy enough

for a prejudiced person to confirm negative opinions.

Another common source of bias in sampling arises from phrasing ques-

tions in ways that encourage certain answers while discouraging others.

Even if a fair sample is asked a question, it is well known that the way a

question is phrased can exert a significant influence on how people will

answer it. Questions like the following are not intended to elicit informa-

tion, but instead to push people’s answers in one direction rather than

another:

97364_ch08_ptg01_177-194.indd 187 15/11/13 10:44 AM

some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has deemed that any suppressed content does not materiallyCopyright 201^3 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights,

affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.