I.

Introduction

Public Opinion, attitudes, perspectives, and preferences of a population toward events, circumstances, and issues of mutual interest. It is characteristically measured by the sample survey or public opinion poll.

II.

Opinion Formation

Public opinion is shaped both by relatively permanent circumstances and by temporary influences. Among the former are the ideas that characterize the popular culture of a given place at a given time. In the U.S., for example, the youth-oriented culture of the late 20th century affects the attitudes of many people toward aging and the elderly. Other fairly permanent circumstances such as race, religion, geographical location, economic status, and educational level can strongly influence the opinions of an individual or a particular group about many subjects. Certain temporary factors also affect the public's attitudes. Among these are the impact of current events; the opinions of influential or authoritative persons; the effect of the mass communications media; and the concerted campaigns of public relations professionals (see Public Relations).

III.

History

The systematic measurement of public attitudes is a 20th-century development. Although occasionally opinion polls were conducted before the 1930s, they were generally neither systematic nor scientific. They dealt with unrepresentative samples or used methods that made certain people far more likely to be included in the poll than others. For example, in “straw polls” the only people counted were those who volunteered to take part.

Public opinion polling improved vastly in the 1930s when business and educational organizations began to develop methods that allowed the relatively unbiased selection of respondents and the systematic gathering of data from a wide cross section of the public. By present-day standards these polls were crude, but their results were in some ways useful. Among the pioneers were George H. Gallup, Elmo Roper, and Archibald M. Crossley.

Two events encouraged polling agencies to further refine their methods. In 1936 a poll conducted by the Literary Digest incorrectly determined that the Republican candidate, Alf Landon, would win the U.S. presidential election. The error arose largely because of biases that caused wealthy people to be overrepresented in the poll. In the 1948 election, most polls mistakenly predicted a victory for the Republican candidate, Thomas E. Dewey, over President Harry S. Truman, again because poor people were underrepresented and also because the polling agencies missed last-minute changes of attitude among the voting public. Since 1948 techniques of public opinion research and polling have improved considerably. Efforts are now made to select respondents without bias, to improve the quality of questionnaires, and to train able and reliable interviewers.

IV.

Uses

Opinion polls are generally accepted as useful tools by business, political organizations, the mass media, and government as well as in academic research. Hundreds of public opinion polling firms operate around the world. Best known in the U.S. are the organizations, such as the Gallup Poll and Harris Poll, the findings of which regularly appear in major newspapers.

In business, polls are used to test consumers' preferences and to discover what it is about a product that gives it appeal. Response to commercial polls aid in planning marketing and advertising strategies and in making changes in a product to increase its sales (see Marketing).

In politics, polls are used to obtain information about voters' attitudes toward issues and candidates, to put forward candidates with winning potential, and to plan campaigns. Polling organizations have also been successful in predicting the outcome of elections. In addition, by polling voters on election day, it is often possible to determine the probable winner even before the voting booths close.

Newspapers, magazines, radio, and television are heavy users of public opinion polling information, especially political information that helps to predict elections or gauge the popularity of government officials and candidates. The public's attitude toward various social, economic, and international issues is also considered newsworthy.

Governments use opinion polls to tap public sentiment about issues of interest. In addition, government agencies use polling methodology to determine unemployment rates, crime rates, and other social and economic indicators.

Polls have been employed extensively in academic research, particularly in the social sciences, where they have proven valuable in studying delinquency, socialization, political attitudes, and economic behavior. Among the prominent organizations that primarily serve academic research purposes are the Survey Research Center at the University of Michigan and the National Opinion Research Center at the University of Chicago.

V.

Methods and Techniques

Public opinion polling involves procedures to draw a representative sample of the population under study. If, for example, one is studying the attitudes of all adults in the U.S., the survey organization would seek to draw up a list of the entire adult population of the country and then select at random a sample to be surveyed. When proper techniques are used and the sample is large enough—1000 to 1500 people—the results obtained are likely to be very close to the results one would get if the entire population were surveyed. Thus, if 60 percent of the sample says it approves of the president's policies, statistical theory shows that if the entire population were surveyed, the probability is 95 percent that between 58 to 62 percent of the people would express the same approval as the sample. The criterion of excellence in a sample is representativeness, not size.

Sampling is vital to the validity of an opinion poll. In practice, however, sampling can be a complicated procedure involving a great deal of estimation and guesswork. The population to be surveyed usually cannot be precisely enumerated. Efforts must be made to break down the population into sampling units of approximately equal size. A certain amount of interviewer discretion is necessary, and complications arise when a proposed respondent is not at home, has moved, or is unwilling to be interviewed. Often only about two-thirds of the intended respondents are actually interviewed and give valid responses. When mail questionnaires are used, problems of nonresponse are higher.

Great care must be used when fashioning the questionnaire or interview schedule, and testing the questions before using them in the field is always advisable. Ideally, questions should be short, clear, direct, and easily comprehended. Apart from such an obvious necessity as trying to avoid bias, many subtle problems arise in framing a question. A word or phrase, for example, may mean different things to different people. In making a question simple enough to be understood by everyone, the issue may be so oversimplified that it has no meaning to the more sophisticated respondent. Sometimes the order in which questions are asked can affect the response. In addition, the tone or wording of the question may alter the measured response: A study once found that the percentage of the public in favor of “forbidding” speeches against democracy was 16 points lower than the percentage in favor of “not allowing” such speeches.

Other problems can be traced to interviewer effects. Age, sex, class, or racial differences between the respondent and the interviewer can sometimes affect the respondent's answer.

Once the opinion data have been gathered, the analyst must seek to find meaning in the results, keeping in mind the problems of sampling variability, question-wording biases, and interviewer effects. The results are tabulated and analyzed using various statistical techniques to determine patterns. Much successful analysis involves comparison: comparing subgroups of the population as they react to the same question; comparing the results of surveys conducted at different times to discover opinion trends; and comparing the responses to different questions. Because of comparison-based analysis, the survey often is divided into small subgroups for comparison—for instance, educated, politically active women with educated, politically active men.

VI.

Criticisms of the Research

Criticisms of public opinion research come from a variety of sources. Many people simply are not convinced that the opinions of a small sample of the population are a viable representation of the opinions of the whole. On this matter, however, the polling agencies can point to the science of statistics and also to decades of experience from which it can be shown, for example, that the same question asked on two different sample surveys at the same time will almost always generate similar results. Other criticisms deal with sample procedures that, for reasons of economy or expediency, sometimes use outdated population data or make compromises with rigorous statistical requirements.

Even assuming that the basic poll data are valid, analyses of the data may be casual and superficial. In some cases, the raw data are simply presented as the public's “opinion” on an issue without deep and careful analysis to probe nuance and possible bias. The subtle influence of variations in question wording on the measured response is often ignored.

In the political area, criticism sometimes focuses on the appropriateness of opinion polling, rather than on its validity. It is argued that elected officials may be too willing to act on what a poll says their constituents think rather than deciding the issues on their merits. Some experts believe that polls may influence voters to favor certain political candidates who seem to be enjoying a notable popularity at the moment. The information that a certain candidate is far ahead in the polls may discourage people from voting at all or encourage them to vote for that candidate and thus may affect the results of the election.