As currently understood, “polls” refer to techniques that combine some form of questionnaire with advanced statistical techniques to produce measurements of public opinion. Although informal straw polls were already used to give rough measures of political support in the latter part of the nineteenth century, they did not gain widespread use until the twentieth century. Political writers such as Walter Lippmann, and social scientists such as Floyd Allport, dissatisfied with subjective descriptions of public opinion, began to look for more rigorous methods of measuring what the public thought of an issue. Social scientists and journalists began to look upon polls, which could be standardized and which also allowed for a more or less accurate measure of public opinion, as a way to solve the need for objectivity in discussions of public opinion.

Despite some notable failures—including an infamous poll run by the Literary Digest for the 1936 presidential election, which seriously misread the level of support for the incumbent Franklin D. Roosevelt—the poll became a nearubiquitous element in public discourse. Although the Gallup organization is still probably the most famous polling group, numerous other firms have developed reputations for their expertise. The Nielsen company’s polls on television viewing patterns, for example, provided perhaps the single most important measurement for most of the American advertising industry in the second half of the twentieth century.

At the same time, the increasing use of the polling data has provoked criticism. Critics of modern political journalism argue that news reporters rely too much on “poll-driven” stories, especially during elections. Political news thus is reduced to stories about which candidate is ahead in the polls, and the strategies used either to get the lead or retain it; discussion of actual policy proposals is pushed aside. Others argue that as a measure of public opinion, polls are essentially misguided, since what they are measuring is in fact an aggregate of various private opinions. More critical theorists point out that polls offer only a limited number of responses to questions, effectively ignoring—and thereby marginalizing—opinions that are too radical or that cannot be standardized. Even social scientists who use polls extensively admit that they present measurement problems, especially the under-reporting of socially deviant behavior and the over-reporting of sociallysanctioned actions.

These criticisms have provoked changes to polling procedures in recent years.

Although polls are still essential to election-year coverage, many news organizations have become more circumspect in their use of polls to drive political news. Among social scientists there is a move to supplement polling with the use of focus groups and interviews, or through the development of “deliberative polls,” in which the polling questionnaire is supplemented by in-depth interviews and interaction between members of the subject population. It remains an open question, however, whether polls as such “measure” or in fact “create” public opinion.