Political Polling and Public Opinion
Wilson Chapter 5















The Market Place of Ideas
Opinions are the products of individuals personalities, social characteristics, and interests.
Opinions are also shaped by institutional, political and governmental forces that make it more likely that
citizens will hold some beliefs and less likely that they will hold others.
In the US and other Western democracies, opinions and beliefs compete for acceptance in what is called
the marketplace of ideas.
Prior to the nineteenth century, each of the various regional, religious, ethnic, and economic groups
generally possessed their own ideas and beliefs based on their own experiences and life circumstances.
The members of different primary groups had little contact with one another, and they knew remarkably
little about the history, customs, or character of their countrymen.
– Transportation problems and the unavailability of effective communications media, and especially in
Europe, an absence of a common national language, kept the city separated from the country.
– Language barriers could be a definite problem. Parisians traveling just a few days from the capital
often reported that it was impossible to understand the language of the local populace.
– Class was also a problem. Members of different social classes, even if they lived near one another
lived in different worlds. Often each class spoke its own language, adhered to its own religious
beliefs, and different beliefs about the political and social universe.
Maintaining the “Marketplace” of Ideas
In the 19th century, the autonomy of the various regions began to diminish.
Western nations engaged in intense efforts to impose a single national language upon their citizens.
In the US, massive waves of immigrants meant that large numbers of new citizens spoke languages
other than English.
– In response the American government as well as state and local governments made vigorous efforts
to impose the English language upon these newcomers.
– Schools were established to provide adults with language skills.
– At the same time, English was the only language of instruction permitted in the public schools.
– Knowledge of English became a prerequisite for American citizenship. With some exceptions, the
efforts of the US and other countries to achieve a common language was successful by the 20th
century.
Maintaining the “Marketplace” of Ideas
literacy - few people prior to the nineteenth century could read and write.
–
Possession of these skills was for the most part limited to the upper class.
–
Widespread illiteracy meant that communication depended on word of mouth.
–
During the 19th and 20th centuries all Western governments actively sought to expand literacy.
–
With the addition of compulsory education, children were taught to read and write the mother
tongue.
The development of communications mechanisms - road building followed by rail and telegraph lines,
radio, telephone, and the satellite based systems of communication of today that link the world have
made communication possible at a moments notice.
Legal protection for free expression of ideas- the West has clearly led the way here.
Maintaining the “Marketplace” of Ideas
Today there are few physical or legal impediments to the transmission of ideas and information across
municipal, regional, class, ethnic, or other primary group lines.
All groups are linked more or less by a common language, mass communication media, and
transportation networks.
The operation of the idea market in the US today has continually exposed individuals to concepts and
information originating outside their own regain, class, or ethnic community.
Because of this steady exposure, every social group has acquired some of the ideas and perspectives
embraced by others. It is virtually impossible to resist some modification of beliefs in this environment.
Political Ideology
 Ideology – a comprehensive and logically ordered set of beliefs about the nature of people and about the
institutions and role of government.
 Ideally, ideological classification would be a convenient way to measure people’s core values and to
summarize their political views.
 Realistically, the result is mixed. People can and do hold a variety of opinions that are influenced by
family, region of the country in which they live and a variety of other factors.
 For the most politically sophisticated, ideological classification goes a long way toward describing their
political convictions. When the less informed answer surveys to determine ideology, the result is much
less indicative of their beliefs.
What Does Liberal and Conservative Actually Mean?
– American politics is dominated by two relatively moderate ideological positions:
 Liberalism-Liberals view the human condition as relative to the quality of society; if we are all
poor the society is a poor one…
– Liberals believe in egalitarianism
– Liberals believe planned change brings the possibility of improved society and
circumstances
 Conservatism- Conservatives view society as a control of humanity’s worst impulses
– Conservatives consider people to be inherently unequal and due unequal rewards
– Conservatives seek order, tradition, and authority
Conservatives are more afraid than liberals of big government, except on matters of law-andorder, and national security
 Conservatives are more likely to see harmful consequences of government help for the
disadvantaged, liberals see the benefits
 Conservatives tend to be moralistic and liberals are more permissive
















Conservative v Liberal
Good indication of the views of liberals and conservatives presented in the themes of the two
ideologies during the 1992 and 1996 presidential campaigns.
Conservatives (Republicans) argued for less government, wrapping themselves in the symbols of
nationalism and patriotism , and took a strong stand in favor of business, the death penalty, and
prayer in schools, all issues popular with most voters.
Liberals, on the other hand wrapped themselves in the symbols of compassion, fairness, equality, and
social justice, also popular issues.
Americans tend to be moderate in views and may not accept the full range of either liberal or
conservative ideology.
Not unusual for Americans to be quite liberal on economic issues and supportive of considerable
government intervention in the economy while holding conservative views on moral and social issues.
Conservatives and Liberals seen to contradict themselves because conservatives want governmentfree life except where moral decisions are concerned – believe in legislating moral behavior and
liberals want government intervention in business and no intervention on moral issues.
Other Ideological Groups in the U.S.
Other groups play minor role: communists, socialists, libertarians
Many parties will be represented in the list of those running for President.
Socialism – based on public ownership of the means of production and exchange.
– Marx described it as the transitional stage between capitalism and communism.
– Under capitalism, the means of production and most of the property are privately owned.
– In communist or socialist systems property is “owned” by the state in common for all the people.
– In the ultimate socialist system, justice is achieved by having participants determine their own needs
and take what is appropriate from the common product of society.
American socialists beliefs: greatly expanded government role, including the nationalization of certain
industries, public jobs programs, much steeper tax burden on the wealthy.
Want to help the underdog by means of income redistribution programs. Would also drastically cut
defense spending.
Libertarianism – emphasizes individual liberty and sharply limited government.
– Carries some overtones of anarchism.
– Has gained a modest following among those who believe that both liberals and conservatives lack
consistency in their attitudes toward the national government.
Beliefs: oppose government and all its programs, favor massive cuts in government spending, an end to
the FBI, CIA, and most regulatory commissions; favor eliminating all government programs not only for
the poor but for business as well.
Would repeal all laws that regulate personal morality including abortion, pornography, prostitution, and
recreational drugs.
Has had a candidate on the ballot in all 50 states in recent elections but never gained more than 1% of
the vote.
Public Opinion: Trustworthy or Fickle?
Top Economic Issue Facing the Country Nearly half of the respondents (47%) who were read a list of
economic issues selected unemployment as the most important economic issue facing the country, while
22% chose the federal budget deficit and 10% cited taxes. Ten years ago, 37% said unemployment, 24%
said the deficit, and 13% said taxes. [Jan 29-Feb 1, 2004; Mar 7-8, 1994]
 The Gender Gap
– Finding Life on Mars: Worthwhile? Men are more likely than women to say it is worthwhile for the
United States to find out whether there were ever living creatures on Mars. Specifically, 51% of men
think this endeavor would be worthwhile, and 46% do not think so. Among women, just 34% think it
would be worthwhile while 66% think it would not be. [Jan 29-Feb 1, 2004]
 The Racial Divide
– Opinion of George W. Bush Just over half of white Americans (56%) say they have a favorable
opinion of George W. Bush, while 43% have an unfavorable opinion of him. Among nonwhites, just
35% have a favorable opinion of Bush, and 64% have an unfavorable opinion of him. [Jan 29-Feb 1,
2004]
The Problem with Polling…
 The main problems of survey measurement, or more accurately mis-measurement, include:
• the sample design (for telephone surveys, how the numbers were selected and how the individuals are
selected within the household);
• the non-availability problem (are people who are available different on the variables we are measuring
than the people who are not available?);
• the refusal problem (is the refusal rate different on the particular variable we are measuring?);
• question wording;
• question order;
• deliberate, or unconscious, lying or false reporting by respondents;
• inappropriate or inadequate weighting of the data.

 In addition, when trying to predict election results, we have several other problems to worry about,
including:
• differential turnout (a huge factor in U.S. elections, where predicting who will vote is often harder than
measuring how they are likely to vote);
• late-swing (i.e., people changing their minds after we’ve surveyed them).
 All of these variables have been shown in various studies to have been sources of error and sometimes
of quite substantial error.
 Unfortunately, we are unable to quantify the effects of these sources of error on our results, or to validate
our results within any kind of reasonable budget.
The Bottom-Line on Polling
 Well designed, well conducted surveys work.
 Their record overall is pretty good. Most social, and marketing, researchers would be very happy with the
average forecasting errors of the polls (less than 2% on two main candidates in presidential races since
1952).
 However, there are enough disasters in the history of election predictions for readers to be cautious
about interpreting the results.
Are We Too Hard On Our Presidents?
The most closely watched political indicator of public opinion is the president’s approval rating.
Presidential approval ratings are thought to be an important component of “presidential power.”
It provides a general guide to the president’s prospects for reelection.
The standard measure is the Gallup poll which regularly asks its respondents whether they approve of
the president’s performance, disapprove of the president’s performance, or have no opinion.
 http://www.gallup.com/poll/stateNation/




Presidential Approval Ratings: George H.W. Bush
 In Jan 1991, after American-led UN coalition forces achieved a swift victory over Iraq, President Bush’s
level of public approval soared to an unprecedented 91%. Many assumed that Bush would coast to an
easy win in the 1992 election. But that of course, did not happen.
 One year later, Bush’s poll standing had fallen below the 50% approval level. By the late summer of
1992, the Bill Clinton had opened a commanding lead in the polls over Bush, who by then could barely
muster a 40% popular approval rating.
The Clinton Years
 After his election too, Clinton found that public opinion could be quite fickle. In May of 1993, only 100
days after the inauguration his approval ratings had fallen.
 According to a May 4-6, 1993, New York Times/CBS News poll, 50% of Americans disapproved with the
way Clinton was handling the economy while only 38% approved.
 The approval rating jumped briefly in June of 1993 when he ordered a missile attack on the Iraqi
headquarters. The attack was in retaliation for an alleged Iraqi plot to assassinate President Bush.
 Public opinion at the end of his second term in office hovered around the 60-65% approval range for
Clinton in spite of his legal problems.
George W. Bush
 The current president’s approval ratings like all presidents’, have fluctuated with the events that have
occurred during his tenure.
 Study the numbers and analyze what has happened during his term.
 http://www.pollingreport.com/BushJob.htm
 http://www.ropercenter.uconn.edu/cgibin/hsrun.exe/Roperweb/PresJob/PresJob.htx;start=HS_fullresults?pr=Bush



The Gender Gap in Political Opinion
The gender gap refers to differences between women and men in political attitudes and in voting
choices. Although the gender gap has historical roots, political differences between women and men
have increased in scope and shown greater persistence in recent years.
Gender played a major role in providing a margin of victory for President Bush. According to the
nationwide election day voter polls conducted by Voter News Service (VNS), a 10 percentage point
gender gap was evident in support for the Republican candidate, with 53% of men and 43% of women
voting for George W. Bush. Similarly, a 12 percentage point gender gap was evident in support for the
Democratic candidate, with 54% of women and 42% of men voting for Al Gore.
The gender gap in the 2000 presidential race was of the same magnitude as the 11 percentage point
gender gap in the 1996 presidential race and much larger than the 4 percentage point gender gap
evident in the 1992 presidential race.
 Men and women were clearly divided in their preferences for president, with a majority of women (54%)
voting for Al Gore and a majority of men (53%) voting for George W. Bush.
 There was a gender gap among white women and white men of the same magnitude as for women and
men overall. White women were evenly divided between the two presidential candidates (48% for Gore
and 49% for Bush) while white men overwhelmingly preferred Bush to Gore (36% for Gore and 60% for
Bush).
 Working women overwhelmingly preferred Gore to Bush (58% for Gore and 39% for Bush).
Presidential Approval Ratings
 All presidents enjoy some time above 50% approval; and most have unhappily spent some time below
the 50% baseline.
 Presidents Truman, Nixon, and Carter all spent time below the critical 30% approval level.
 Three sources account for most variation in a president’s approval rating:
–
–
–
The “honeymoon effect”
The “rally round the flag effect”
The effect of ups or downs in the economy
Approval Ratings:
The Honeymoon Effect
 Every president starts the term with a rather high level of political support, with approval ratings in the
70% range not unusual at the beginning of the term.
 Then as the president’s term unfolds, the approval rating undergoes a gradual but inevitable decline.
 It is not surprising to find that presidents start out with an aura of goodwill, with even supporters of the
defeated opponent offering their approval.
The Honeymoon Effect
 Why does the honeymoon fade?
–
–
Presidents cannot please all segments of society all the time.
 Some supporters will eventually be upset and this disillusionment creates a spiral of declining
approval
 The longer they government the lower the approval rating seems to be the general rule-unless
unexpected events occur.
A second interpretation is more optimistic.
 A president’s early loss of popularity is best considered as a retreat from a starting point that was
artificially high at the onset, rather than an erosion of a natural base.
 About a year into the presidency, the rating reaches a normal level.
 By the second year, the president’s popularity level will converge toward the middle 50% range.
 From this point on, it is as likely to rise as to fall with events impacting the numbers.
 The basic reason why presidents start out with an artificially high approval rating is that newly elected
presidents start out free of criticism-either from politicians or the media.
– This is because the president has proven popularity at the ballot box.
 As the presidential term evolves, however, the president must take actions that are subject to secondguessing by the public and the media and the numbers will fall or rise accordingly.
Rally-Round-The-Flag Effect
 From time to time some foreign policy event will occur that will impact the president’s approval rating.
–
–
–
–
The Cuban Missile Crisis, 1962
Seizures of U.S. ships by communist foreign governments in 1968 (U.S.S. Pueblo by the North
Koreans)
Nixon’s Moscow Treaty in 1972
Carter’s Camp David Accords, 1978
 These types of events are typically followed by a surge in the approval rating of the president.
 Going to war is a special case.
–
In the short term, wars traditionally result in a showering of approval on the president.
 Roosevelt’s entry into WWII caused a surge in his ratings from the mid 50’s to the 70’s.
 More recent wars show similar patterns.
 The biggest gain from a war was George Bush’s popularity surge to Gallup’s record high of 87%
in March of 1991 from 59% in January of the same year following the brief Gulf War against Iraq
 George W. Bush’s rating also surged following 9/11 Terrorist Attack to 90% on Sept. 22, 2001
from 57% in August.
 The political rewards from wars and invasions can be short-lived.
–
–
–
–
The Johnson administration miscalculated the public’s tolerance for the prolonged Vietnam War.
Johnson fell from the eighties to the thirties in a few short years.
Truman declined during the Korean War even further into the 20’s.
Both Truman and Johnson declined to seek reelection.
Bush’s approval showed both what a war can do for popularity, but also how fleeting it is as his
numbers fell to 46% by January of 1992.
The Economy
 It’s common knowledge that the president’s approval rating rises and falls with the economy.
 When unemployment or inflation rise, the president is blamed. When these indicators rise the president
is praised.
 The electorate seems to read the state of the economy as a sign of the president’s competence.
 The exact nature of this conversion is uncertain, but seems in its simplest version to be a straight-forward
response to the personal “pocketbook’ considerations – when people face good economic times they
support the president and in bad times they don’t.
 Most people however know that they do not actually owe their current economic fortunes to the
president.
 But people do recognize the state of the general economy as having some relevance to their personal
situation.
 People evaluate presidential performance on future performance, not past performance, usually. This
means the most current state of the economy is the most important to the president’s numbers and
reelection.
Public Opinion Polls
Public opinion polls are the most efficient way to determine the opinions of a range of people, not
just those who are vocal.

Polls:
–
indicate the direction, the stability, and to some extent the intensity of public opinion at the time
the poll was taken.
–
reveal which groups of people are most likely to support certain policies, be concerned about
certain issues, and vote for particular candidates
–
indicate which issues do not generate much public interest or awareness – may not need to be
addressed by public officials
–
provide feedback to candidates about voters’ concerns, political attitudes and their reactions to
symbols, issues and images that the candidates are using.
 Polls enable candidates to assess the damage caused by an opponent’s charges almost immediately
and respond to them quickly.

– The failure to do so can be fatal: examples being Michael Dukakis in 1988 when Bush charged that
he was soft on crime, weak on defense and out of touch with American values.
– In 1992 and 1996 Clinton used polls to anticipate and reply to charges against him in less than 24
hours.
 Polls cannot predict what future opinions will be. Those taken several days before an election may not
accurately forecast the results if a large number of voters make up their minds after the polls are
completed.
– Examples: Harry Truman predicted to lose in 1948 and Ronald Reagan’s win in 1980
underestimated.
 Polls are only a snapshot that policy makers may consider; are not a detailed guide for action















Political Poll History
The first straw poll – 1824 by the Harrisburg Pennsylvanian to assess public support for the four
candidates seeking the presidency.
Called straw polls because they are the equivalent of tossing a piece of straw into the air to detect
the direction of the wind. Polls sample a small group of people to determine the attitudes of the
electorate as a whole.
First presidential prediction made in 1892 when the New York Herald tallied polls collected by
newspaper editors across the nation.
The Literary Digest Polls
Largest and most comprehensive of the early national surveys: conducted by the Literary Digest.
Mailed millions of ballots and questionnaires to names of automobile owners and in telephone
directories.
Conducted this type of poll in the 1924, 1928, and in 1932 elections and correctly predicted the winner
but in 1936 during the height of the Great Depression, it predicted a huge win for the Republican Alf
Landon instead of Roosevelt.
Instead Roosevelt won with 60%of the vote.
Major flaw: questionnaires received did not represent the public at large, few at that point could own
either cars or phones. They had become semi-luxuries.
The Birth of the Gallup Poll
George Gallop – believed that his desire for direct democracy called for public information that was not
filtered through the economic elite.
Therefore he tested his belief•fs accuracy in his mother•fs election for local office in Iowa.
Shortly thereafter he began predicting the results of elections for dissemination by the public media.
His polling group still makes predictions today and have been joined by many others.
Polling Across The Decades
The 1950’s – most politicians have little interest in political polling. Some polling was done but was
mostly to get a “picture” of where the campaign stood rather than being a central part of the strategy.
The Harris poll begins and one of first clients is Sen. John F. Kennedy. Harris continued to work with
Kennedy through the 1960 primaries. Harris is credited with spawning the new industry of political
polling.
The 1960’s – private political polling began to take hold. Private means that candidates paid pollsters
to be on their staff for the duration of the campaign.
– it’s during this period that political polling became a line item in the budgets of many races.
Though included, it did not have the influence on strategy that is has since obtained.

















Campaign Polling 1967-78
During this period there were only about 3 or 4 national polling firms conducting polls for each major
political party. This was due to:
–
large infrastructure and network needed to conduct the type of polling used in those days
–
the market was not yet producing enough dollars to support many additional research firms.
The pollster and polling function still was to provide a unique set of numbers and data to the media
consultants and direct voter contact functions in the campaign. The polls were not really being used
as part of the campaign strategy, only used as a means of determining harm by other candidates and
response to previously determined strategy.
Method of polling used: in-home, face-to-face interviews; very expensive and time consuming and not
as accurate as the telephone polling which would evolve later.
By the Carter – Ford campaign, both candidates were using telephone polling
Campaign Polling 1979-1988
telephone polling became the norm
more frequent polls were possible
all polling became more affordable so even local races could poll
Both PACs and the major political parties frequently commissioned polls
Independent polling firms began to spring up due to the drop in cost of doing the polls
Campaign Polling 1992-Beyond
beginning with the 1992 Clinton/Bush campaign polling became central to campaign strategy.
Information is a valuable tool in determining on what the campaign should focus and the best
approach to take to make the candidate more electable.
The pollsters role is to provide input in two critical questions in the campaign:
– What is the best message for this campaign?
– What is the definition of our key target group of voters?
Other questions answered by the polls include –

What are the best channels to use in communicating with our target voter?

What intensity or emotionalism should we use in communicating with the target voters?
All American electoral contests today use some type of political opinion research technique to gauge the
mood or collective mindset of the electorate. It’s essential to success.
Most campaigns with more than 10-15,000 voters and more than $50,000 in campaign dollars to spend
will use some type of professional, modern research technique to guide them in developing and/or
refining their campaign strategy
Virtually all gubernatorial and U.S. Senate campaigns rely heavily on the use of polls. The cost to their
campaign of these frequent polls – 5-10% of the total campaign budget
Everyone else, U.S. House campaigns and local campaigns are often calling on polls to give them an
edge but the size of their campaign budget may limit their use of polls.
What Does the Pollster Do?
 The Early Phase of the Campaign:
– the benchmark poll
 pollster attempts to become immersed in the campaign itself – interviewing staff, fundraisers,
friends of the candidate
 becomes familiar with the candidate’s voting record if any, family and background, and staff as
well as party officials, the opponent and his/her staff
 benchmark poll is long and time consuming to complete (90-100 questions) usually conducted
with a large sample (500-1200 people in statewide races, 400-500 in congressional races)
What Does the Pollster Do?
 Mid-campaign – usually the longest period in the campaign
–
–
three roles of the pollster:
 constant advisor or “reality check” on all campaign decisions which might have an impact on
voters
 a monitor – through trend polling, determining the success of the tactics being used
 evaluator – of different methods of communicating the goals of the election, from the stump
speech to the television spot
trend polls similar sample size to the benchmark poll, but the questionnaire is shorter
What Does the Pollster Do?
 Endgame – four to eight weeks to go in the campaign
–
–
–
–




candidates are firing their best shots at those voters still persuadable as well as firing at each other
between 40-50% of voters today report that they make up their mind for whom to vote in October or
within 5 weeks of election day
key role for the pollster is to determine who these people are
trend is for more frequent use of polls especially in the days just preceding the election
How Are Polls Conducted?
A representative sample of voters much be selected; it is essential that they are representative of the
population as a whole
random sampling – means that everyone in the population has an equal chance of being selected for
the survey
General rule: the larger the sample the more accurate it will be. Not always the case but more often
than not it is.
When assessing the results of the sampling two factors must be considered:
–
sampling error – the degree to which the sample could deviate from the population as a whole. If
a poll with a sampling error of + 3 reported that 50% of those surveyed thought the president was
doing a good job, pollsters could generalize that anywhere from 47-53% of the entire population
held that opinion.
–
Most national surveys are based on a level of confidence of 95% meaning that 19 times out of 20
the results would be within its sampling error.
The Dangers in Polling
 Bias can be introduced in the survey by using emotional or controversial words
– For example terms such as “pro-choice”, “baby-killing”, “rape” or “incest” in a question on abortion will
affect the response.
– So do phrases like “don’t you believe?” “isn’t it true?” and “most people believe, don’t you?” Even
slight changes in the wording can produce large differences in repsonses.
 Questions of the agree/disagree variety may be biased toward the agree response.







The choice of open- or closed –ended questions may affect the kinds of responses obtained.
– Closed-ended questions force a person to choose among a list of responses
– This can cause people to express opinions on issues about which they have little or no information or
that they do not feel strongly about or they may force responses into categories that do not
accurately reflect the respondent’s opinion.
– Open-ended questions have no predetermined answer
The order in which the questions are asked can affect the responses people give.
– Information included in the early questions may be used to answer later questions, particularly if
people try to be consistent in their answers.
The interview itself can affect the response.
– An interviewer who develops a rapport with the respondent may be able to elicit more information
than can be obtained from an impersonal questionnaire.
– Also the interviewer can encourage responses that are considered socially acceptable or desirable
and suppress those that are not.
Finally the interpretation of the results may be flawed.
Interpreting Polls
Check the sampling error.
– If the margin of error exceeds the margin separating the two responses, one response cannot be
interpreted as more probable that the other.
– For example if a random sampling error of + 4 % and Candidate A leads Candidate B by 2% say
51% to 49% then the contest is too close to call.
Determine whether the group sponsoring or conducting the poll have a vested interest in the
results.
– Treat skeptically any finding disclosed by candidates, their organizations, and their supporters,
including interest groups
Examine the survey questions.
– Do the questions measure what the pollsters claim they do?
 Examine the survey questions.
– Do the questions measure what the pollsters claim they do?
– Remember the wording of the questions, their placement in the survey, and responses permitted can
all affect the results. Be sensitive to any positively or negatively valued buzzwords, symbols, goals,
or names of individuals that may influence responses
– Check the time frame which the poll was conducted. Have any events that might change
respondents’ opinions occurred between the time the poll was taken and the day the results were
published?
 Remember that random selection is the key to sampling accuracy.
– Discount the local media polls that ask people to call in or mail in their responses. These only
measure the preferences of people concerned enough to take the time to respond.