|
|
|
The Value of Public Opinion
Polling
A Post-Mortem of the 2003 Knoxville Mayoral
Election
Dr. Michael M.
Gant
Telephone surveys have become a ubiquitous feature of contemporary
American culture. Whether in the form of marketing surveys, public opinion
polls or candidate surveys, everyone seems to want to know what we think.
For some people, participating in surveys can be a welcome opportunity to
express their opinions; for others, it can be an irritating and unwelcome
intrusion.
Reading
or hearing about survey results likewise produces varied reactions. For
some, polls are a source of information about what other people think. For
others, survey results are a source of amusement. For still others, poll
results can be quite upsetting, especially when a poll suggests that those
surveyed hold different opinions than those held by the reader. As Director
of the Social Science Research Institute at the University of Tennessee,
Knoxville, I have found this last reaction not only to be the most common,
but also the most difficult to answer.
SSRI is heavily engaged in sample survey research. One of the most
important components of our work is a series of state-wide public opinion
surveys known as The Tennessee Poll. When the results of our polls are
released, I can always count on hearing from folks who refuse to accept the
findings. Among the arguments I hear are the following:
1. Some people reject the idea that scientifically surveying several
hundred respondents allows for meaningful statements about the entire
state.
2. Some argue that if we had just asked the question differently, we would
have had different results; usually, these suggestions would bias responses
toward the critic's position.
3. Some are convinced that since the University of Tennessee
is involved, the survey must be rigged to advance the University's
interests.
4. And some will argue that, since they don't feel that way, and since no
one they know feels that way, we must be wrong.
When it comes to attitudes and opinions, it is very difficult to convince
critics that poll results are correct. However, there are some polls that
can be checked for accuracy-election polls. Since such polls always report
candidate preferences, these results can be checked against the outcome of
an election to determine how accurate the poll was. As a case in point,
consider the recently contested municipal elections, when Knoxville elected it's
first new mayor in sixteen years. Bill Haslam
defeated Madeline Rogero in a surprisingly
competitive campaign. I say "surprisingly competitive," because
many local observers expected, and still others explicitly predicted, the Haslam campaign to roll over Rogero.
The final margin of about six percentage points seemed to catch a lot of
people off guard. The day after the primary, Haslam
campaign manager Bill Lyons was quoted in the News Sentinel as saying,
"I didn't think it would be this close."
These expectations arose from several sources. First, while Madeline Rogero is certainly not an unknown, many assumed the Haslam family name would give him a sizable advantage.
Second, Bill Haslam amassed a campaign war chest
fully four times as large as Rogero's. And Rogero's support was widely seen to come
disproportionately from socio-economic groupings that traditionally vote in
lower numbers than those supporting Haslam.
Continued
on Page 2
|
|
