Polls on the military’s “Don’t Ask, Don’t Tell” policy reported very
diff erent fi ndings, making it diffi cult to know what the public really
thought. The policy was eventually repealed, enabling gay men and
lesbians to serve openly.RESPONDING TO PUBLIC OPINION ON GAYS AND
LESBIANS IN THE MILITARY
As we discuss in Chapter 9, elected offi cials in America work
hard to cast votes and take other actions that their constituents
will like. At fi rst glance, this behavior seems easy. All a politi-
cian needs to do is take a poll, measure public opinion in her
state or district, and comply with the demands expressed in the
survey responses.
Actually, it is not easy for a representative to fi nd out what
her constituents want, because public opinion is hard to mea-
sure. Everything you have learned in this chapter suggests
that taking a poll doesn’t necessarily tell the whole story. Poll
results need interpretation and may not provide clear guidance
to elected offi cials.
Consider public opinion on allowing gays and lesbians to
serve in the military. Suppose a member of the House repre-
sents a district where public opinion on this issue is the same
as the national data in Table 5.2 (p. 137). Also assume that this
representative wants to mirror district opinion to gain political
support and stay in offi ce. What sort of guide does the survey
data provide to the representative?
The fi rst problem is that the survey provides different guid-
ance depending on which question the representative looks at.
As we discuss in this chapter, small differences in question
wording can produce large changes in responses. Our repre-
sentative could fi nd data showing support for several different
interpretations—that her district is sharply divided or that it is
in favor of changing the policy.
These data reveal two problems with reading public opinion
for guidance. In many cases, opinions are sensitive to question
wording, so a representative cannot be sure that survey results
are indicative of actual feelings or are an artifact of how the ques-
tions were asked. Moreover, since opinions are typically formed
on the spot, based on relatively little information, even if the sur-
vey questions were not changed, opinions might look very differ-
ent if a survey was taken a day, a month, or a year later, as people
take account of new or different information. The problem is not
that the people who conduct surveys try to bias their results—
rather, the diffi culty in measuring public opinion stems from how
Americans think about politics and respond to survey questions.
For these reasons, voting in line with polls may not be politi-
cally advantageous. A representative who did so might fi nd later
fi nd that the poll’s results were shaped by question wording,
and that her constituents actually preferred a different policy
and a different vote—or that opinions shifted between the time
she voted and the next election because people changed how
they formed their opinions on the issue. And, as we discussed
earlier, when opinions are changing, as they are in the case of
gay marriage, a poll taken at one point in time may not be a
good guide to public opinion a short time afterward.
These fi ndings create a quandary for elected offi cials: even if
they want to vote in accordance with constituent opinion, these
opinions are hard to measure and may change over time for a
variety of reasons. But representatives must vote, even when
they are not sure what their constituents want. How should they
vote? You decide.You Decide
SOURCES OF OPINIONS| 133Critical Thinking Questions
- Looking though this chapter, do you see cases
where you think opinion surveys provide an accu-
rate picture of public opinion? - What advice would you give representatives on the
limits of polling and how to read survey fi ndings?