Paper prepared for the plenary session at the annual meeting of World Association for Public Opinion Research (WAPOR) in Prague, Czech Republic September 17-19, 2003.
Check Against Delivery

Social Science Theories and the Study of Public Opinion

Hans L Zetterberg[*]


A first aim of this document is to bring to the attention of pollsters and journalists an integration of some major social science theories relevant to the interpretation of public opinion. A second aim of the paper is to overcome the discrepancy between polled opinions, as we meet them in today’s media, and public opinions, as we meet them in the established theories of social science. Public opinions of both kinds are expressed in language. Different conceptions of “the meaning of meaning” lie at the core of both the theory of opinions and of the interviewing practice in the study of opinion. The inspirations from George H. Mead’s symbolic interactionism and Ferdinand de Saussure’s linguistic structuralism are noted. We then review public opinions in various social settings: (1) opinions in encounters showing their motivating force as discovered by Charles H. Cooley, William James, and Émile Durkheim; (2) opinion variation in positions and roles revealed by demographic background factors and summarized in Robert K. Merton’s concepts of status-set, status-sequence, and role-set; (3) opinions in organizations, networks, and media as developed by Robert E. Park and The Chicago School of Sociology. At this juncture we list the difference and similarity between (a) polled opinions rooted in encounters and recorded by interviews in random samples, and (b) public opinion of classical theory recorded by examination of sources in archives and rooted in lasting and functioning networks. We continue with (4) opinions depending for their survival on media as discovered in Noelle-Neumann’s spiral of silence; (5) opinion content as revealed by Mary Douglas group-grid classification; (6) opinions set by stratification in society, primarily by the central zone as described by Edward R. Shils; (7) public opinion as national creed in Gunnar Myrdal’s sense; and (8) ideologies and opinions among the makers, keepers, brokers, and takers of welfare and other cardinal values, according to the author’s own theory.

Public Opinion

In the Spirit of Rousseau

In classical works of social science, public opinion is usually conceived as a property of a collective, an expression of the collective’s conception of itself and its role in history. Individuals could have a more or less correct interpretation of this volonté générale, to use Rousseau’s term. If their interpretations were too deviant, they became viewed as being stupid, unaware, false, insane, or, at worst, criminal and a danger to society.

But Rousseau could give another meaning to public opinion, volonté de tous, the will of all. In this case, opinion is an attribute of an individual, not of the collective. It can be questioned and discussed, and it can be summed up as majorities and minorities.

Eliminating the metaphysical qualities of volonté générale but otherwise following the lead of Rousseau, opinions divide into two well-known categories:

·         opinions we must express in order to be in a collective and represent it to others, and

·         other opinions that we may express without being disliked by, isolated in, or exiled from our collective.

The first category provides a basis for determining the consensus of opinion, and the second a basis for determining the differentiation of opinions.

The amplification of the latter has led to two phenomena: on the one hand the mechanisms of decision-making by casting votes in parliamentary assemblies and general elections, on the other, opinion polls and the use of their results in journalism. The former leads to binding decisions in the form of legislation. This process of majority rule attempts to approximate John Locke’s ideal of a government based on the consent of the governed. The latter leads to knowledge about how people think and feel, and is appropriately called "opinion research." The two are entirely different, but may, of course, develop links to one another.

The extent to which politicians in various settings use the findings of opinion research in their legislative work is an empirical question with different answers in different cases. Usually elected politicians pay attention to public opinion polls particularly at election time, but their modus vivendi is to follow their local or regional traditions, their own personal convictions, the party platform, discussions at party caucuses, leads from party activists, reports from government agencies, expert testimony, suggestions given by lobbyists, and information from mass media. A recent review and two case histories from The United States have been provided by Jacobs & Shapiro (2000) with the telling title Politicians Don’t Pander.

Public opinion has relevance for several areas of social science. Among them are opinions that inspire, motivate, and control people in their everyday life, opinions about common issues for people and about the resolution of these issues, opinions that give legitimacy to governments and other social institutions, and opinions that characterize subgroups, periods or regions.

As scholarly concerns, all these topics  antedate the advent of opinion polls, a recent social invention that had its breakthrough in the United States in the 1930s. There was much knowledge and several theories about these topics prior to polls. As in all humanistic research, this knowledge was based on source material in archives, diaries, letters, memoirs, newspapers and other published material, minutes of debates, etc.

With polling opinion researchers used interviews with questionnaires to create their own source material. The quality of this new kind of source material in social science rests on the questions asked and on the selection of respondents.

Creating Source Material for Opinion Research by Means of a Questionnaire

Let us assume that a survey reseracher knows the techniques and the arts of asking survery questions, as they are summarized in books such as Sudman & Bradburn (1982), Yurangeau, Rips & Rasinski (2000) and Noelle-Neumann & Petersen, (2000). This makes the opinion researcher an expert on everything needed, except the topic of the research at hand. The first rule for questionnaire authors is then to know the topic of their study and to find the various indicators of this topic. The serious writing of a questionnaire begins, therefore, with a trip to the library, or a search in a data base, or with an interview with experts on the reserach topic. If omitted, we may get embarrasing results. For example, every fifth question in a survey of the Panama Treaty 1977 contained factual errors (Smith & Hogan 1987).

Responses to questionnaires are source material, not conclusions, for the opinion researcher. The questionnaire must provide the researcher with all the information needed to reach the solution to a research problem. It is the opinion researcher who shall calculate the answer to the research problem, not the respondents. The contributions of the latter are indicators in the form of responses to simple questions.

A full-fledged pre-election poll is thus based on questions giving indications of turnout, party and/or candidate preferences, and climate of opinion. The full-fledged study of opinions about a new issue may need a series of auxiliary interview questions in addition to the query how the respondent views the issue. For example, to know the quality of the answers the pollster may have to ask: Had you heard about this issue before I asked? Have you noticed the issue being discussed on television, in papers? Have you searched for information about the issue? Have you talked to others about the issue? In what context did you talk about the issue? Did the others in such talks have the same views about the issue as you, or other views? Did the conversation cover the consequences of the various views on the issue?

The questions to be answered by the research project are something very different from the questions in the interviews. This principle is discussed in the survey literature, lately and most definitively by Noelle-Neumann & Petersen, (2000, pp 93-95). The sponsors of polls and the editors who publish them do not always understand this principle. What is worse, too often pollsters themselves get trapped into putting a research question straight into their questionnaires instead of translating it into a series of interview questions. When answers then appear in brief slots in TV newscasts, the effect is not necessarily information about public opinion. The effect is rather a premature closure of the process of opinion formation; the audience is given the impression that the issue is already settled (Yankelovich 1991).

The Language in Questionnaires

The author of a questionnaire for a survey of the general public must use an everyday language common to the entire population. Slang, newly fashionable words, and esoteric words are obviously ruled out. But so-called ordinary language has also become a rare commodity for the questionnaire writer. Jürgen Habermas (1981) defined one context he calls Lebenswelt, ‘life-world.’ The life world of daily activities is imbued with traditions from many generations in rendering accepted interpretations of symbols. In the life world their meanings are self-evident. Not so in the System World of modern institutions. Modernization and globalization reduce the sway of the shared meanings of the life world. Mankind's present disorientation in the universe of meanings is a price paid for the diversity occurring when a modern world of institutional systems colonizes the life worlds. In later writings Habermas has explored the opportunities for democratic discourse to overcome such difficulties of diverse meanings.

Specialization has also reduced the realm of common meanings. The French philosopher Jean-Francois Lyotard (1993) studied the consequences of the absence of a common language in advanced, differentiated societies. Nowadays no person needs to understand very much about areas in society other than his or her own. Our words and meanings are determined by the contexts of different life spheres and cannot be adequately understood outside of them. Academic research, politics, economics, sports, art, literature, and religion, all have their own symbols and languages. No area in life commands a pan-language. A modern man must maintain skepticism towards metanarratives.

My brief experience as an editor-in-chief of a metropolitan newspaper confirmed Lyotard's view. Our many-splendored society is like a newspaper. In its different pages or sections, a big daily paper mirrors and helps define and redefine society. The pages or sections have different editors, affectionately called “space barons” by a far from sovereign editor-in-chief. Each section of the paper has its own criteria and makes its own evaluations about what is worth publishing. No space baron possesses criteria that are applicable to all the others.

However, the writing of questionnaires to the general public is different from journalism. In one respect it is  like tabloid writing – short words, short sentences, short paragraphs – but without any sensationalism. A second rule of questionnaire writing is to use language common to all life spheres of society. This makes the writing of questionnaires for opinion polls more difficult than ordinary writing of nonfiction.

The further study of questionnaire writing requires us to dwell on “the meaning of meaning.” A ‘symbol’ – like a word used in a questionnaire – is that device by which we on any one occasion can represent an image and/or a notion and use it in conversation with others. This definition hints that there are two kinds of symbols: those related to images and those unrelated to images but found in other notions used in social interactions. I shall call them Meadian and Saussaurian to honor two great scholars of the study of symbols.

In the spirit of the young Wittgenstein we can say that Meadian symbols depict something in the same way as pictures do. A symbol may thus depict reality or fantasy, something present or absent, something in the past or in the future. Its meaning is the image it conveys. The Chicago philosopher George Herbert Mead’s posthumous book Mind, Self and Society from 1934 became the foundation for a school of thought about society called “symbolic interactionism.” He distinguished between gestures and significant symbols. A gesture is a part of a behavior sequence that signals the total sequence, for example, a dog, baring his teeth and assuming a certain posture is a gesture meaning "fight" to another dog. A symbol can be a gesture that evokes the same meaning in the receiver as it does in the transmitter: "in this case we have a symbol which answers to a meaning in the experience of the first individual and which also calls out that meaning in the second individual.” The person who cries "fire!" to his neighbor shares images of what is going on within him. In fact, both the one who has seen the actual fire and the one who has only heard the shout "fire!" react in similar ways; for example, by escaping or by starting rescue work.

G. H. Mead tended to overlook that all symbols do not refer to something fixed outside them. The very relations between the symbols can also define their meaning. This idea had been launched already in 1916 in a classical book by the Swiss linguist Ferdinand de Saussure, Cours de linguistique générale. His starting point is “the arbitrariness of the sign”. There are no particular relations between the sounds of the mouth pronouncing the word child and its meaning. The word does not look like a child, crawl and walk like a child, cry and laugh like a child. But in the part of mankind that has learned English it means child. So do the words kid and youngster. Symbols that can replace one another in a number of presentations have the same meaning; symbols that are irreplaceable in presentations have unique meanings. I shall call such meanings "Saussurian.”

Ferdinand de Saussure inspired a school of thought about society called "structuralism.” Claude Lévy-Strauss became its leading social scientist. In his theory of society and culture, symbols are the active agents engaged in a great struggle of survival. In such a theory, man is rather incidental, a mere accessory that helps certain symbols in their struggle for survival and hinders others.

A group of leading French scholars in the late twentieth century trained in structuralism turned against Lévy-Strauss. Jacques Derrida and other “deconstructionalists” drew the ultimate conclusion of an exclusive use of Saussurian meanings and their iterability. If symbols get their meaning only from their place in presentations, meanings may shift from time to time in an arbitrary way. Language, literature, legislation, education, and everything else involving symbols, are then mere games: chaotic games with ever-shifting rules.

Michel Foucault used this conclusion to deny that there were any objective truths. What people talk about as true statements do not tell us how things really are but about who is in charge and has the power to establish the meanings of our symbols.

Outside of France such views were called “post-structuralism.” It has some elements that are empirically grounded. It cannot be rejected simply by denouncing it as a fitting ideology for nihilists and anarchists. The search for Saussurian meanings is at the bottom of much scholarship in contemporary social theory and cultural and literary criticism.

A more viable critique of poststructuralist views focuses on the fact that languages also have meanings in the form of shared and stable images. George Herbert Mead is needed to rescue us from a chaotic abyss of post-structuralism in which social scientists, journalists, and critics of culture lose all bearings.

There is a lesson here for pollsters. A third rule of questionnaire writing is to avoid as much as possible words with pure Saussurian meanings, i.e. those established by assuming the interchangeability of symbols. Legal phrases, bureaucratic and technocratic languages abound with Saussurian symbols. So does much literary and artistic language. Such language cannot readily be used in questionnaires to a broad public. By necessity, questionnaire writing becomes void of sophisticated and lovely garlands of words. Pollsters instead fill their questionnaires with words that have unambiguous and untarnished Meadian meanings, i.e. depict something in the same way as pictures do.

To make the critical ideas of difficult subject areas easily understood, the polling questions might in fact be supported by pictures, a practice used and refined by Elisabeth Noelle Neumann in Institut für Demaskopie Allensbach in Germany. Meadian words – preferably supported by pictures – allow us an effective use of questionnaires in complex, differentiated societies far beyond Habermas’ narrow life worlds.

The Structure of Questionnaires

The writer of a questionnaire must know not only the ins and outs of the topic of his research and the restrictions of questionnaire language, but also how to structure a questionnaire. The fourth rule for questionnaire writing is to give it the structure of a conversation, albeit one with a series of short question-answer sequences. A good questionnaire has an introduction with a few simple questions. They are followed by general questions, then specific ones, as in a funnel. Sensitive and difficult questions are usually saved until typical respondents have warmed up to their task. The order of the topics in a questionnaire should be such that it minimizes the influence of previous questions on the answers to later questions. The questionnaire should give each reasonable response alternative the same chance to be expressed. It is unacceptable to lead the respondent to choose an alternative that is formulated as socially more acceptable, or one that is biased by the preferences of the questionnaire writer and/or his client. Of course, the acceptable questionnaire also asks about no more than one issue at a time, so that one knows exactly which issue an answer refers to.

The sequence of questioning in a questionnaire is further guided by so-called "filter questions." They define subgroups in the sample, e.g. eligible voters to be asked about their voting intention, or car owners to be asked about their plans to replace their car. Those who are not caught by the filter are not asked such questions.

Most questionnaires in opinion research set up a conversation in which the respondents get used to giving short pre-proposed answers: Agree-Disagree, Approve-Disapprove, More-Less, Right-Wrong, Yes-No. If such a sequence of questions is interrupted with an open-ended question you cannot expect a full answer rich in nuances. To get good answers to open-ended questions they must come in sequences of their own – as in an in-depth-interview – when respondents are prepared to answer in their own words.

An Encounter Known as The Interview

When there is a high likelihood of cognition about another person's actions we speak of an 'encounter' (Goffman, 1961, pp. 8-11). The sum total of actions in all encounters a person has may be called his or her 'repertoire of action.' This repertoire grows with the number of encounters we have with others who differ from the ones we met before. It grows with education and travel, with the reading of fiction, with the films and plays we see. In general, city life entails more encounters than rural life.

The encounter is the broadest of all terms used in social science for interpersonal contacts. The encounter between the writer of the questionnaire and a person in the sample, a 'respondent', is the basic unit in opinion research.

The cognitions we have of each others' actions in an encounter need not be scientifically correct. W. I. Thomas' dictum applies here: if people define a situation as real it becomes real in its consequences. For example, if a government believes that a nation prone to hostilities has weapons of mass destruction, this belief has real consequences irrespective of the correctness of the belief.

An encounter does not have to be conceived as a meeting between living people. We encounter Socrates by reading Plato's dialogues. Christians encounter Jesus by listening to sermons and understanding the Gospel. If we have any liberal education at all, we have had many encounters with the great intellectuals of the past through their written words. And we encounter distant persons of our own time, not only through books but through all mass media.

An encounter does not have to be conceived as an actual physical meeting. Some encounters with contemporaries are indirect, i.e. with invisible others. In traditional kingdoms the subjects might never have met the king in whose name they were regulated and taxed. In contemporary society indirect encounters abound (Giddens 1990, 21-29). We do not meet the policeman who signed a parking ticket and affixed it to the windshield of our car. A person can rent an apartment without an appointment with the landlord who signed his lease. And you can get a credit card without ever seeing the person who provided its authorization. You buy things by mail order or Internet without meeting a salesman. In public opinion polling the respondents never meet the author of the questionnaire.

Encounters may be direct, that is, face-to-face, or indirect, that is, without the visible presence of others. The questionnaire can reach the respondents directly by mail, internet, or reprints in media. Or, it may be mediated by an interviewer who meets the respondent face-to-face or by telephone. When pollsters en masse changed to telephone interviewing in the last few decades of the twentieth century they lost the possibility of using pictures; it may now be regained with new generations of cellular phones or with questionnaires on the Web.

An encounter can span time and space, the visible associates and the invisible ones. However, the archetypical encounter is a face-to-face meeting. As laymen we usually generalize experiences from our face-face-encounters to the historical and indirect ones. As scholars of social science we ought to proceed more cautiously and test whether different forms of indirect encounters actually do have the same consequences as face-to-face encounters. It should, for example, not be taken for granted in advance of evidence that the various forms of contact used in public opinion interviewing give the same answers.

On Opinions in Encounters

Opinion and The Motivation To Look out for Ourselves

The uniqueness of an individual is not only a matter of a more or less unique combination of genes but also of a unique combination of encounters, past or present (Simmel 1908, Chap. 6). Our self-image is neither entirely self-made, nor entirely inherited; it is also influenced by the encounters we have or have had. The latter is an aspect that will concern us here for it fuels actions that run in a channel from the social to the personal.

People have a tendency to develop self-images that harmonize with (i.e. are synonymous or consonant with) the public views about them in their encounters.

This is known as Cooley's (1902) hypothesis about "the looking-glass self"; Cooley was the first to formulate cogently the idea that a person's self is a mirror-like reflection of what others think of him, thus discovering that society and personality are, at least in part, two sides of the same coin.

People have an inclination toward those actions within their repertoire of actions which result in maintaining their self-images.

In 1890, before Cooley formulated his theorem, the American philosopher William James had expressed the motivational significance of public opinions in the form of evaluations by others in his famous dictum: "A man's social me is the recognition which he gets from his mates.... Properly speaking, a man has as many social selves as there are individuals who recognize him and carry an image of him in their mind. To wound any one of these images is to wound him" (James 1890, Chap. 10, Section 1b). The James proposition presupposes a more stable core – the conception of oneself – the complex of a human individual, and it suggests that preservation of this core is a factor governing other activities

Both Cooley and James saw a very broad avenue for the social construction of personality here. Nowadays we recognize many other sources of human motivation, many inherited and some acquired. The preservation and development of a self-image shaped by others may not be our strongest motive. Its significance lies rather in the fact that it is the one motive that can be most readily manipulated by others. To give or withhold food, sex, capital punishment is very cumbersome and costly. To give or withhold praise or blame for an action, deference or disdain to a position, is simple and cheap. The opinion of others is an easily used master key to motivation.

The idea can be restated as a conclusion flowing from the combination of the idea of a looking-glass self with the idea of defense of one's self-image:

People have an inclination to choose those words and actions within their repertoire which maintain the public view of them in their encounters.

If we fail to choose those words or actions we feel embarrassed, lose face. The fascinations of public opinion are many, and a most fascinating aspect is its force of social motivation. The ups and downs of “approval ratings” of presidents and prime ministers and other public personalities provided by pollsters are a case in point.

The above Tenet of Social Motivation presupposes Cooley's convergence between the opinions of us held by others and our own self-image. If this were the full story it would be correct to adopt a totally cynical view of human life: everyone is on the lookout for Number One. But in reality it is one factor among many. Even the master of the cynical approach, Irving Goffman, admits that that we also attend to our fellow men and like to preserve our meetings with them (Goffman 1967). So far we have dealt only with the cynical aspects of motivation: looking out for ourselves. Let us now turn to the trustful aspects: looking out for others.

Opinion and The Motivation To Look out for Others

Our motivation to preserve our good standing in our encounters presupposes that the encounters do take place and that they can produce reasonably stable public views. Thus the first motive to choose words and actions to gain approval from others in encounters is soon supplemented by a second motive to choose words and actions that maintain encounters and protect the encounters that reward us.

Some people can claim quite rightly that they do not care much about what people in certain encounters think about them. But there are other encounters where the opinions of others about us are exceedingly important to us. These are the opinions of our “significant others,” to use the term introduced by Herbert Mead. We do not really know a person until we know who his or her significant others are.

Psychologists have pointed out that we attach special importance to how people in our childhood, especially our parents, view us and our behavior. They are our "primary groups," the first in our life to influence us. Old-fashioned parents and authoritarian teachers can equip children with a mental gyroscope that will inform them throughout life whether they are deviating from behavior that would meet with the approval of their parents and teachers.

In the 1950s social scientists recognized that an increasingly large number of Americans wished, above all, to gain the approval of their “peer groups” – others who are similar to them in age, occupation, et cetera – rather than approval from parents and teachers. Riesman (1953) called such people “outer-directed,” in contrast to the earlier “inner-directed” types. The outer-directed lack a gyroscope, but have instead sensitive radars that signal what those around them think.

Riesman's well-known typology needs to be supplemented by significant others from the historic past and significant others from the near future.

Around the middle of the 1700s and for more than a century thereafter leading scholars of pedagogy in German-speaking countries believed that the key to bringing up good citizens lay in antiquity, in the legacy from Athens, Rome, and Jerusalem. Bracing quotations from the Romans would steer youth toward that which is right, true, and beautiful.

The attempt to offer generations of youth significant others from the classics ran out in the sand. Wilhelm I, the nationalistic German emperor, declared in 1890 at a school conference he himself had initiated: “It is our duty to educate young men (sic) to become young Germans, not Greeks or Romans.” Yet there are still some individuals in Germany and elsewhere for whom historical models and ideals remain a reality.

There is also the possibility that one’s significant others can be anticipated contacts in future encounters, or even figures of the imagination. It is not only in fairy tales that a girl may make herself ready for the prince who will come – one day. An assiduous small businessman may adapt his present decisions to what he anticipates that the big bank president or industrial leader would expect – one day. Or, the artist, misunderstood by his contemporaries, may persevere in creating, bolstered by the conviction that critics and the public will appreciate his work and recognize his greatness – one day. In like manner a politician who aspires to become a statesman may anticipate the judgment of future historians. In other words, the opinions of “anticipated significant others” may also have great influence on our opinions. Thus we state about The Maintaining of Encounters:

People have a tendency to maintain rewarding encounters with significant others, be they historical, primary, peers, or anticipated.

An unexpected quantitative confirmation of this proposition comes from survey research.

Explaining Falling Response Rates

The-run-of-the-mill survey interview is not  necessarily a particularly rewarding encounter. To the respondents, the questionnaire writer (and the interviewer, if any) are not very significant others. In the main, respondents do not eagerly look forward to the next interview. Since polling interviews are voluntary, some choose to decline the next one. Thus we get in the survey-rich countries a secular trend of declining response rates.

The widespread use of the sample survey instrument has lead to its increasing uselessness as a provider of reliable statistics about the entire population. When the response rate in the advanced countries over the past half-century has declined from 80 à 90 percent to 50 à 60 percent, polling cannot count on support from mathematical statistics. The present models for calculating the level of a variable and the confidence intervals for this level require high response rates unless the relation between non-responses and the topic studied is known in advance, which is rare. The calculations of “margins of error” in samples with poor response rates do not reveal the range in which the true value is found with 95 percent probability, as is often claimed in press releases. At best it hints that repeats of the samples with the same inadequate response rate in 95 cases of 100 would render values within this range. In this way we may understand why different polls differ slightly from one another in their results, but not how they differ from the true value in the entire population studied.

There is an obvious need for the polling industry in survey-rich countries to make their interviews more rewarding and significant to the respondents.

Opinions and Norm Enforcement

The tendency of people to maintain the existence of rewarding and significant encounters is supplemented by a more profound tendency to enforce the norms that prevail in these encounters.

Why does a majority in an encounter punish a deviating minority by giving them an unfavorable evaluation? In a famous passage from the sociological literature Émile Durkheim argued that punishment had less to do with the criminal but was in the main designed to uphold the moral order of the law-abiding population. (Durkheim 1893, Chap. 2, Sec. 4) Maintaining of encounters, I would say, occurs because people do not want to be traitors to encounters that uphold their self-image. The Punishment Proposition says that people punish the noncompliance they come across so that they can maintain their own orderly evaluation. The implications of this process are perhaps best seen if they are stated in the negative mode:

If persons in an encounter do not give an unfavorable evaluation to those who deviate from a norm in the encounter, they lower their own self-evaluation.

In other words, deviations can only be tolerated at the price of lowering the evaluations received by the compliant participants in an encounter. If a person deviates from a norm but nevertheless retains a favorable evaluation, the entire scale of evaluation has changed so that everyone in the encounter is degraded a notch.

A difficulty for our acceptance of The Punishment Proposition is that, so far as I know, it has not been subjected to any direct empirical test. The main reason we believe in it rests on the fact that its implications seem empirically sound. One such implication, the moral panic, is relevant to opinion pollsters and we shall review it here.

Moral Panics

In the 1960s drugs became a not uncommon part of student life in many Western countries. Some of the social researchers who had had personal experiences of the reactions to the use of marijuana during their student days later used these experiences to pose scientific questions. Why did society react so strongly against drug use? All research indicated that the abuse of alcohol was far more widespread and had far more costly consequences than drugs.

The researchers began to gather material on other social reactions of disproportionate magnitude in relation to the prevalence of the phenomena. They found many: for example, unbalances in reactions to abortions, aids, child abuse, cruelty to animals, homosexuality, incest, food additives, pornography, prostitution, radioactivity, Satanism, trafficking. Analyses of these phenomena led to a theory about moral panic.

Moral panic is a contagious wrath or fear that arises when the behavior of certain people (the wicked ones) is deemed to be so pernicious or threatening to the rest of society (the good ones) that it calls for immediate measures to control the behavior and restore the social order. Leading researchers on moral panic were the South African Stanley Cohen, who had studied the phenomenon in England and coined the term moral panic, as well as Erich Goode in the United States and Nachman Ben-Yehuda in Israel. The latter two wrote a book on the subject (Goode and Ben-Yehuda 1994) and defined 'moral panic' by these criteria:

First, moral panic arises through a distortion of the magnitude of an issue, which appears more ominous than it really is.

Second, moral panic is expressed in an emotional pitch charged with outrage, disgust, despair, fury, protest, and a desire to punish.

Third, moral panic generates a belief among those involved in its creation that danger threatens. In a community in Norway, the belief that incest was a widespread occurrence was able to unite disparate elements of society.

Fourth, moral panic creates a divide between the “good” and the “wicked,” an opposition that can be exploited by interest groups, political parties, and religious groups.

Lastly, moral panic demands that measures be taken, and it often echoes in parliaments. Moral panic appears to be the voice of the people.

A list of examples of moral panic has certain resemblances to tabloid headlines. A moral panic is newsworthy of itself and should, of course, be reported in the media. Sometimes, however, the editorial staff is itself caught up in the throes of the panic and becomes more than a purveyor of news: it orchestrates the panic. The editorial and news pages run the same campaign, TV joins in and turns up the volume, the waves of indignation mount ever higher. Pollsters are called upon to document its prevalence. In such situations, democracy, like any other form of governance, is in danger of turning into mob rule.

A moral panic can then also erupt in the government and parliament. Politicians react personally with indignation as everyone else does, but they are also pressured into action by the media, the polls, and public demonstrations. In the 1990s a moral panic in regard to pornography using children broke out in sexually tolerant Sweden. It captured the royal court and large parts of the government, and threatened to subvert the country’s central and hard-won constitutional principles of the rule of law and the freedom of the press.

On Opinions in Positions and Roles

As language is used in encounters, some parts of it are frozen into elementary social forms. The process may be called “mundane structuration.”

A man who has done much fishing one day shows a young boy in his community how to fish. He tells him and demonstrates the ins and outs of fishing, saying ”Do this!” and “Do that!” Another day he shows another boy how to fish, and a third and a forth. In the course of these events he becomes known in the community as “fishing teacher.” The language in the community has in this way structured a position. Others may take on the position of “fishing teacher” along with him, or after him. And others take on the position of "fishing apprentices". The community now has an established relation of teacher-apprentice in fishing. Any talk about learning how to fish or any other action showing others how to fish is henceforth both enabled and constrained by this structuration.

Many social scientists have found that a comfortable vocabulary when talking about structuration in society is found in the language of drama: that is, roles, publics, scenes, actors, and so forth. The behaviorist language gave way to a dramaturgic language. This trend started with the social anthropologist Ralph Linton (1936, pp 115ff) who made "status" and "role" cornerstones in describing social structures. I will rephrase his terms, and then show how a later theorist, Robert K. Merton, developed them and applied them to the analysis of opinions.


When they are speaking generally, sociologists and anthropologists use the term position (some say “status”, others “identity”) to include every capacity in which an individual can be expected to act. Typically, the grammatical subject of a prescription is a word that defines ‘position.’

Students shall go to class.
Gentlemen are requested to wear jackets in the dining room.
Drivers should proceed carefully.
The First Mate is responsible for loading the cargo onto the ship.

The names at the left are descriptions of persons, not their proper names but designations of structured categories of actors. The crucial aspect of these descriptions is that they are subjects in prescriptions. Each one defines a position.

Positions may be self-described. Some participants in an abortion debate claim to speak for the unborn child; this does not mean that they have been designated to do so by unborn children. Some in the globalization debate claim to speak for the Third World; this does not mean that they have been nominated by Third World populations. Self-designated positions become significant only to the extent they are accepted by others. An impostor claiming to be the legitimate heir to the throne of Russia can get a limited following. If self-designated positions are not accepted by others we deal with people such as the tragic inmates in mental asylums who think they are President of the United States.

When we ask “Who is she (or he)?” the first routine answer is to mention one or more positions. “She is a student, lives in a suburban house, married, has a child, and has a part-time job downtown.” The list of positions a person currently holds is what Merton (1957, pp 380-384) called “status-set.” They represent his or her current ‘commitments.’

As a second answer to the question “Who is she?” we may add that when she finished high school she did not immediately start college. She held a couple of good jobs and started a family. When she later in life wanted a college education she took a part-time job and entered a School of General Studies at a nearby university. The list of her past positions constitutes what Merton calls her “status-sequence,” that which in everyday parlance is called her ‘career.’


To identify a 'social relation' (or “role”) we mention two positions and hyphenate them: parent-child is one social relation; student-teacher is another; customer-dealer a third, and so on almost ad infinitum. Technically speaking, two positions define a social relation if the prescriptions addressed to one contain references to the other: "Children should obey their parents"; "Parents should guide their children."

Occupants of the two positions that constitute a social relation are referred to as 'associates' and the prescriptions involved are 'role prescriptions' (some say role expectations). It is often useful to indicate which social relations a person can enter into by virtue of occupying a given position − e.g., student-professor, student-student, student-dean − and his or her associates in any social relation. Information in these areas, ‘the role-set’ in the terminology of Merton (1957, p 369), constitutes a third routine answer to the question, “Who is she?”

Accounting for Opinion by Background Factors

The main demographic base of all social science are the numbers of men and women, young and old, and their division, if not in terms of race, which no longer is supported by biology, so according to birthplace or ethnic origin. In recent years many social scientists have redefined the categories of sex, age, and origin by using concepts such as ‘gender,’ ‘life course,’ and ‘ethnicity’ to bring out their social content.

The notion of a status-set gives theoretical relevance to the demographics included in every opinion poll. A main rule of thumb in opinion research is that people of the same age, sex, occupation, marital status, et cetera have similar opinions. The general principle is:

Persons with similar status-sets, status sequences, and role-sets have similar opinions.

If some with similar status-sets nevertheless do have different opinions, we use a second rule of thumb, namely that they have gone through different status-sequences. Here, class, ethnicity, religious affiliation, among other factors, make an impact. Here we also find the most important enduring influence that education has on opinion (Hyman 1975). If people with similar status-sets and status-sequences still have widely different opinions, we apply a third rule of thumb that says it is likely that they have different role-sets, that is, they are influenced by associates representing different views of the world.

While these steps go a long way to account for variations in public opinion, they do not go all the way. There remains a large residual of more personal opinion formation. This residual of status and role theory may actually represent the most dynamic aspect of human living (Cicourel 1974), namely active individuals’ constant search for cues, defining and redefining situations, their pursuit of knowledge, riches, power and culture, and their elaboration of their lifestyles.

On Opinions in Organizations, Networks, and Media

The concept of encounters has served us well, particularly in the study of daily face-to-face interactions. We have seen their structuration into positions and roles. Now we shall focus on how positions and roles combine into still larger and more lasting communication structures. As positions and relations are formed, they cluster into organizations, networks, and media, all more lasting structures than encounters. Our starting point will be taken from ideas in the school of sociology that emerged at the University of Chicago in the first half of the twentieth century. We begin by updating the ideas to fit current terminology.

A founding member of the Chicago School of sociology, Robert E. Park, had studied in Germany, and his doctoral dissertation in 1904 at Bonn University had the title Masse und Publikum. A "mass," he said, is an agglomeration of people without contacts with one another but exposed to a common source of information, e.g. the same newspaper. A "public" is a gathering in which people talk to one another and become aware of another’s viewpoints. This was the beginning of a schema of different structures of social interaction that came to mark the early teachings of sociology: the group, the public, the crowd (Park & Burgess 1924).

To streamline this thinking, we note that lasting forms of interaction can be distinguished in part by the reciprocity of contacts and in part by the existence of a shared source of communication (leadership). Using these two dimensions we can define communication structures. The combinations provide four types. All are clusters of interconnected positions and roles. See Figure 1. Our definitions give different labels to the participants and to the structures in which they are participating. This streamlining of the terminology means that Park's original definitions and labels are changed.

Figure 1. Overview of Lasting Communication Structures

Is there a common symbolic environment?




Yes or No

Is there a common
source of communication?





Are there mutual channels
of contact?











'Group members'




Illustrations from

Schools, Universities
Academies, Institutes

Learned societies, Seminars

Lectures, Textbooks, Journals, Scholarly works



Firms, Coops, Trade organizations. Unions, Cartels

Markets for goods, services, raw material, money, patents, etc




Lobby groups

Juntas, Parties
Election campaigns



4. ART

Theaters, Museums, Cultural institutions

Coteries for dance, play, visual arts, literature



5. RELIGION Synagogues, Churches, Mosques, Temples, Congregations Sects,
Sacred texts
6. MORALITY Guidance agencies, Institutions for sick, elderly, small children Charities for donating time & money,
Moral movements
Appeals to idealism & ethics  


An 'organization' has a common leadership with designated positions and other defined positions that are filled by coworkers (or staff) and members with different positions. Contact between the different positions is maintained through established channels. Typical examples are a government agency with department heads and their staffs, a business concern with a managing director, middle management, and other personnel, an association with a chairperson, functionaries, and rank-and files.

The actors in organizations may all be called ‘members.’ The standard measures taken of members are their numbers and ranks.

The Staffs of Organizations

'Administrative bureaucrats' are the agents of the leadership's abiding desire to govern and keep control of an organization. They include civil servants in routine jobs, the ombudsmen of organizations, and the staff of corporations. These bureaucrats advance arguments that strictly follow the goals and rules set up by their superiors. In many countries they may threaten refractory elements with recourse to the judiciary, the police, or the military. This bureaucracy has been called the "reptile brain" of organizations (Gouldner, 1979, p 50). It regards freedom of opinion as a disruptive element at the workplace.

'Technical bureaucrats' or 'professional technocrats' are different. In addition to the rule book of the organization, they have specialized knowledge relevant to their jobs, usually acquired at colleges or universities. In modern society they administer production, health care, and communication. They characteristically make the organizations they administer more efficient and shape them so that they get a greater say in them than the purely administrative bureaucrats. They seldom own the organizations, but nonetheless keep them under a kind of control. Inspectors of nuclear power plants or radiation experts do not own the plants, but by virtue of their professional competence they can, in practice, decide how the plants are to be run.

Freedom of Opinion in Organizations

In contrast to administrative bureaucrats, the professional technocrats justify their actions not only by referring to the instructions from their superiors or the resolutions of congresses but also by pointing to facts and reason. A professional technocrat wants to be able to criticize that which he considers to be superstition, be it in regard to the dangers of the use of pesticides, the availability of investment capital, or the health risks of using prophylactics, etc. He then becomes a champion of freedom of opinion. The technocrats are dependent on a certain kind of freedom of opinion, namely the freedom of appeal to rationality. With its help they can free themselves to some extent from the powers-that-be and from the administrative bureaucracy.

This freedom of technocrats has been pretty well established in modern states. In authoritarian or totalitarian states, however, the process has regularly been threatened by one of the characteristics of Stalinism: the subordination of the technocrats to administrative bureaucrats.


In a 'network' (some anthropologists say "grid") people maintain contact with one another in an informal way, but the network lacks a common and defined leadership position. It does, however, have foci. Networks may actually come into existence by those facing a common issue, for example, by residents in a building complex who have not previously known each other but join together in a network to clean up their neighborhood after a storm.

Early writers about public opinion such as Lawrence Lowell (1913) saw the public as a network, a huge discussion group, producing public opinions. This was in line with the conception of an Öffentlichkeit as analyzed by Habermas (1965). Since ancient and medieval times public opinion has been formed at market places, or, in the aftermath of religious gatherings, or at similar occasions.

Networks have participants, not members as do organizations. They are ‘publics.’ The standard measure of a network is ‘size,’ that is the number of persons involved, and its ‘density,’ that is, the ratio between actual contacts and potential contacts that exists between the participants.

Networks are fluid parts of a society. They fill the space between organizations. They follow the rules of sociability, not the rules of leaders or authorities. They may mix members who are familiar with each other with some who are strangers. The glue that keeps a network together is the issues faced by the network.

Free markets in the economy are networks; they share no common leadership that decides the price and volume of exchanges of goods and services. But the participants in the market have issues to focus on: the refrigerator to purchase, the transport to provide, and so forth. Voters may exchange political views while meeting for coffee; the issues may be a candidate’s competence or a clause in the party platform. In the literature on democracy, agora, the marketplace of ideas, has become a common metaphor for the phenomenon.

Informal networks almost always also arise also within organizations. You get only a limited understanding of how an organization works by studying its formal plan of positions, departments, division of responsibility and accountability. Knowledge of its informal organization, its internal networks, is necessary for a full understanding of how things operate. A new, externally recruited executive needs time to get a good grasp of his new organization. The fact that we have so much informal networking in organizations means that we can study public opinion inside organizations. Pollsters are often called upon to do employee surveys.

In the area where networks border on organizations we find the ‘sect.’ One usually speaks of sects in a religious context, but a sect need not be a religious community. It can be a marketplace for vegetarian food, a political movement, an aesthetic coterie, the proponents of a scientific theory. The word ‘sect’ means followers.

A sect is a community of equals, held together by a firm, common standpoint. Its leader is only a spokesperson for its followers, not the leader of a staff as in an organization. Spokespersons can easily come and go, and the internal life of the sect can be very fulfilling for its members. Here one can find an abundance of emotional warmth and involvement. Here one finds like-minded fellowship. Here one finds equality. Here there is no bureaucracy.

The big problem for sects is to keep its followers together. You can understand sects only if you realize that one of their overarching needs is to keep its network intact. This need infiltrates almost all of the sect’s activities, even those that are apparently unrelated to it. Putting a priority on internal cohesion means an avoidance of very concrete answers to practical questions. It is enough to be against sin, apartheid, the atom bomb, gene manipulation, the ozone hole, etc. A detailed program and constructive compromises with adversaries might only lead to disagreement and schisms.

Many sects become missionary, and strive to spread their views and fellowship to others. A missionary position obliges everyone who wishes to remain in the community to acknowledge that they desire to convert outsiders. Of course it benefits the sect to convert heathens, which is the manifest function of its missionary efforts. But it also benefits the sect to reinforce the sense of solidarity among its members, which is the latent function of its missionizing. As noted, a primary aim of a sect is usually the cohesion of its network.


A 'medium' is a communications structure with a common source (pulpit, stage, editorial board) whose communication is one-way. TV, newspaper, theaters are typical examples.

The rapid expansion of media in the last century affected the conception of public opinion. For example, Walter Lippmann (1922) emphasized public opinion as a property procuced by mass media. For a period prior to the event of opinion polls, the notion of public opinion was virtually synonymous with opinions expressed in pamphlets and editorials.

The actors served by media are ‘audience.’ The standard measures of a medium are ‘reach,’ that is, how many persons are in its audience, and ‘frequency,’ that is, how often the persons are in contact with the medium. A whole industry financed by advertisers and publishers has emerged to record the shifting reach and frequency of popular media. Its enterprises are sometimes merged in the same organizations as market research and opinion polling.

The print media and early ether media – books, newspapers, radio, TV – are straight one-way communications. The recipients consume them without influencing content. Their power over them lies only in the choice to stop reading, listening, and viewing.

Interactive media – the new electronic media of cyberspace – undo part of the recipient’s connection to passive consumption of a text or image. The recipient no longer sits staring at a screen; to an ever greater extent he enters into a dialog with it. He chooses the program, he chooses alternative courses of action in computer games, he participates in debates in a virtual discussion group, and he participates in decisions about the ending in so-called interactive stories. Interactive media enable him to break out of the role of passive observer of dramas written by others and to actively take part, not only in the drama itself, but also in setting up the rules it is to follow. The role of interactive media in opinion formation is a virgin research area; here we await a canonical text.

Editors and The Journalistic Process

You cannot report everything. The journalistic process is one of never ending selection. The wire services present long menus. Editorial meetings are held to decide which stories to pursue. The reporters who are given assignment to pursue stories must sift and sort facts and views in the raw material that they can muster. The subeditors select among the stories available what shall be published, what shall be rejected outright, what shall be reworked, shortened or strengthened, and what shall remain in store for another day. The board of directors has sifted and selected candidates for an editor-in-chief, who, together with colleagues, has sifted and selected and appointed various subeditors. Editorial policies have been hammered out from many alternatives; options have been rejected and adopted. In this long chain of picking and choosing lie also the details and colors that make a news story come alive. The noble art of journalism is to make sure that life and reality remain when the journalistic process comes to end at the moment of publication.

There is a German tradition of journalism that regards the ultimate task of the journalist to be to convey a Weltanschauung — a philosophy of life. (Karl Marx, of course, worked in this tradition.) There is a British tradition of seeing the function of journalism as providing a good story — that is, news or a report with intrigue, drama and a sense of immediacy. There is an American journalistic tradition whereby the primary task is the exposé, or "muckraking" — exposing evils, misuse of power and the like. Journalism is not immune to the fickleness of fashion: for some time (since Watergate) the muckraking tradition has been held in high regard.

Public opinion makes itself felt also in the editorial offices and the studios of the mass media. It vibrates easily through the typically open landscape of news desks. It affects the process of news selection and presentation. The amount of self-censorship is considerable. (Zetterberg 1992)

Research into the opinion climate of editorial offices would more than anything I can imagine illuminate the mechanisms of media power. What we have so far of penetrating insights into this milieu comes, not from research, but from thorough American court proceedings in connection with libel trials such as Westmoreland versus the Columbia Broadcasting System and Sharon versus Time (Adler 1988)

The editorial office is actually an organization coupled to a media structure. So is the business office, with the managers and owners of the medium. The scholars of the Frankfurt School developed a critical theory that the latter set the tone of the media content (Adorno 1991). In the main, however, these scholars ignored the fact that media studies need to deal with so-called autopoietic subsystems, soon to be discussed. Media owners may disturb editors, but cannot run their daily tasks.


All the above communication structures presuppose a common symbolic environment. For a 'mass' this condition need not apply. However, the term "mass" has been used in so many ways in social science that it is virtually useless (Bell 1956). Park applied it to what we nowadays call media and their audiences. Ortega y Gasset used it about people who have incompetent judgments about the efforts required to have a rich modern life. Others have used it to characterize modern society. For example, the followers of Lewis Mumford or Jacques Ellul, use the term to designate a dehumanized society based more on machines and attendants of machines than on free people. And Karl Mannheim applied it to the overly bureaucratic society.

‘Mass’ can perhaps still be a useful concept if we stick to the idea that it means undifferentiated encounters of many atomized persons with awareness of one another but without communication with one another, without any common issues to focus on, without a leadership to guide and coordinate actions.’ One example would be the milling travelers at a big international airport, a lasting structure, repeated day after day but with different participants. Another example would be travelers in underground trains who see each other but do not speak with each other. But a mass may give those involved common experiences, such as the discomfort of crowding, or generate common reactions, such as bursts of laughter or panic attacks.

The Structuration of Issues

Whatever an organization, a network or a part of a network, or a mass medium focuses on is called 'issues.' Thus we have organizational issues, network issues, and media issues. And we have scientific, economic, political, artistic, religious, and ethical issues. Figure 1 guides us not only in sorting out structures but also in sorting the multitude of issues.

Some issues may grow into 'social movements,' a situation where the same issue concerns a number of organizations, networks, and media. These movements seem to have a natural life cycle. Combining the insights in an article by Downs (1972) and a book by Spector and Kitsuse (1977) we can identify six stages.

  1. Pre-problem stage. A harmful situation exists and is observed in some encounters but has not yet attracted the attention of lawmakers, journalists, or the public. Small networks and groups make initial claims and begin to recruit support.
  2. Alarmed discovery and euphoric enthusiasm. A dramatic event creates larger public support for solving the problem. Circular reactions in demonstrations and manifestations have a mobilizing effect.
  3. Official recognition of problem. Established leadership gets involved: there may be legislation or a creation of agencies to deal with the harmful situation.
  4. Gradual decline in public interest. Recognizing the costs of significant progress and becoming bored with the problem, media attention fades, and the public loses interest. At this stage the movements will survive only if they find new issues.
  5. Active dissatisfaction in original groups. The groups who made initial claims reemerge and express dissatisfaction with how the harmful situation is being handled. Some who have lost confidence in how the problem is being handled try to create revivals with broader or more radical agendas.
  6. Post-problem stage. In spite of the fact that only limited improvement has been achieved, the issue at hand is replaced by new problems.

These stages are useful benchmarks in assessing public reaction in modern societies to potentially harmful problems. Knowledge of these stages is also useful in assessing the impact on social change of various movements, o on efforts to lengthen the life of a movement.

A sophisticated observer does not take for granted that a social movement will have its day and blow over. The century-old environmental movement has switched its concerns from issues of conservation and national parks to any recreation on lakes or in mountains and woods, to abatement of noise and congestion in cities, to air and water pollution, to abandoning nuclear energy, to the clean-up of poisonous waste, to biological diversity, to an agriculture free from chemical fertilizers and pesticides, to global warming. A core of ecology has remained in all these changes. The shifting foci have given environmentalism a much longer life than a typical social movement.

The Interplay of Organizations, Networks, and Media

Organizations, networks, and media impinge on the individual in different ways. Organizations define positions and identities in more clear-cut ways than do networks. In an organization your position is defined by the organization. In a tight network of neighbors this may also be true, but in a less dense network of relative strangers you can define your identity with considerable discretion. You can more easily present yourself in terms of past and present positions of your own choosing.

More important is the fact that the effect of communication on individuals is much dependent on the combination of their organizations, networks, and media. Only some of these combinations have so far been topics of penetrating research and theorizing.

Crossing Organizations and Networks

The longstanding difference between the British chartered corporation (organization) and the market economy of the Manchester School (networks), inspired anthropologist Mary Douglas to formulate her so-called cultural theory. She shows that the structural differences between organization and network result in different cultures. Her cross-classification of group and grid has gone through several modifications. Figure 2 from Fardon (1999, p 225) shows some of the labels and synonyms used by Douglas between 1978 and 1996.

Figure 2. The Group-grid and Culture According to Mary Douglas





Atomized subordination
Backwater isolation

Strong group
Central community
Ascribed hierarchy
Conservative hierarchy
Collectivism with structure

Competitive individualism
Active individualism

Dissident enclave
Egalitarian enclave
Egalitarian collectivism

Many consequences arise from this typology, a fact that has made it very useful in social science. Of special interest to opinion research is that it locates the structures supporting some major political parties: agrarian, conservative, neoliberal, and egalitarian parties (Wildavsky 1971, 1991)

Crossing Media and Networks

The recipients of media communications can have reciprocal contacts with only a few other recipients or none at all. An important form of social influence goes from a person with frequent and multiple media consumption to his networks of persons with much fewer contacts with the same media (Katz and Lazarsfeld 1955).

People who are well integrated into their networks, be it colleagues in the informal conversations at lunchtime at their workplace, neighbors, friends, relatives, et cetera, are, in all likelihood, adept at rapidly picking up what others think. This enables them to give candid expression to their version of public opinion and to perhaps a few odd aberrant personal opinions as well, without disturbing the others in the network. They know public opinion whether or not they have been exposed to the media.

Those who are poorly integrated into their network, and therefore lack knowledge of what are serviceable opinions in conversation with relative strangers, run more of a risk of "putting their foot in it.” Newspapers, radio and TV, rather than personal contacts, become their principal source of information on what others believe and think, i.e., which 'climate of opinion' prevails. For them, the media alone define the nature of current conventional wisdom and teach them public opinion. From this perspective, the media teach the public not so much what to believe and think, but rather which opinions are acceptable in public intercourse. Thus media have their greatest impact where networks are weak, a point originally made by Park.

When the public reach for their newspapers or other mass media, seeking news, entertainment, and potential subjects of conversation for the day, support for their own interests and views is not always forthcoming. Instead, they find the journalists' selection of subjects and views. Failing to find their own opinions in the newspaper gives pause to many readers. They may even lose their self-confidence and withhold their own views in everyday conversations. What they really believe then fall into a "spiral of silence" (Noelle-Neumann 1980). Media are not at all omnipotent in creating public opinions, but in the struggle for survival among opinions, the media play a more decisive role.

Classical Public Opinion and Modern Polled Opinion

Some classical theorists of democracy tended to imbue public opinion with awesome powers – the voice of the people speaks to them like a voice of God in the old days. I am not the only pollster in the world who has gratefully acknowledged the aura given to my work by much democratic theory, but have  privately been doubtful whether the simple responses I reported from the field matched these theorists’ weighty conceptions of public opinion.

Already in the 1940s, scholars became concerned whether polled opinion is public opinion as conceived in social and political philosophy. For example, some sociologists of the Chicago persuasion with their well worked-out theory of networks, organizations, or mediums accepted polls of participants in functional networks that could discuss public issues, but they rejected national polls of random and scatterd respondents (Blumer 1948). Alternatively, they suggested national polling practices to better meet considerations of their theory (Rose 1949).

Half a century later, many tensions between social science theories and the practice of polling remain unresolved. Pollsters have developed their techniques and methods to great sophistication. Social and political theorists have developed more profound visions of democracy and of government by discussion. (For an overview, see the thirteen essays in Bohman & Regh 1997.) But methods of polling and theories of public opinion still live apart. The two can be bridged only by special efforts involving agony for both parties.

There are many similarities and differences between the classical humanistic approach to opinion study in archives and libraries and modern practices of polling in the field. Here is a list that begins with relative similarities and that ends with the major difference.

·         Where there is no known issue – i.e. focus for an individual, an encounter, an organization, network, or medium – there is no opinion. Opinions unrelated to known and discussed issues rarely turn up in the archives used by researchers in the humanities. In polls that create their own archives they turn up more often. The reason is that pollsters often miss the first acid test of the existence of an opinion about X, namely the first essential filter question in public opinion research "Have you ever thought about [the issue of X]?" The value of a response to an issue you never thought about until a questionnaire or an interviewer brings it up is limited, and should perhaps not be called public opinion.

·         An 'individual opinion' consists of the cognitions (descriptions), attitudes (evaluations), and exhortations (prescriptions) concerning an issue. Individual opinions are habitual views; we repeat them as the occasion arises. An individual may change his or her opinion over time, but totally haphazard expressions do not qualify as opinions. The humanist ideally looks for more than one expression in an archive before attributing an opinion to a person. The pollster should ideally repeat or slightly vary his question to a respondent to see if he gets the same answer (reliability testing). However, this ideal that both subscribe to is only rarely turned into practice.

·         A 'collective opinion' consists of the beliefs (descriptions), sentiments (evaluations), and norms (prescriptions) concerning an issue. They are customary views; we find them among a multitude. Multitudes can be counted, and we may talk of majorities and minorities. Pollsters have an advantage here. A branch of mathematics, the theory of sampling, has become an intellectual basis of survey research, of which opinion polls are an example. The general tenet of statistical sampling is simple: by studying the response in a randomly selected sample of the population we can, with known probability, estimate the response of the entire population. We often use a similar approach in daily life. A cook need not consume a whole pot of sauce in order to know how it tastes. He need only stir the sauce (randomize) and taste a spoonful (sample) of it. If the sample has too little spice, the whole sauce has too little spice. This can be concluded with a known degree of conviction expressed as statistical confidence limits. The statistical "margins of error" do not mean that that the researchers have done something in error, but rather that the readers or users of the findings would be wrong in assuming that the reported numbers are exact representations of reality rather than indications of likely intervals. Unfortunately, as we have seen, the response rates to surveys have declined, and this makes statistical testing less meaningful.

·         Collective opinions may be hidden or public. 'Hidden opinions' are only expressed in close, neighborly relations, or in the privacy of a voting booth. What you write in a private letter or diary, or express to your family in the seclusion of your living room is not necessarily public opinion. The term 'public opinion' refers to collective opinions about an issue expressed in public, i.e. views freely expressed also to relative strangers. The recording of public opinion in the old days was restricted to the analysis of speeches and deliberations open to the public, and to any material – pamphlets, columns, editorials, et cetera – published to the public. The polled opinions of today are also public opinion since they are expressed to a relative stranger, i.e. the interviewer or the questionnaire writer. But very often, pollsters fail to check the public nature of the responses by the second essential filter question “Have you ever talked about [issue X]? Of course, the quality of a response to an issue you never have talked about until the questionnaire or interviewer arrives is limited.

·         Public opinions are often stereotyped. The stereotyping of opinions is the extent to which a shorthand label is repeatedly used in place of the full language that normally required to express the opinion. Stereotyped actions take less of our time and energy than the ones adjusted to the details of the situation. Walter Lippmann, who brought the issue of stereotypes to general attention, says:

"There is economy in this. For the attempt to see all things freshly and in detail, rather than as types and generalities, is exhausting, and among busy affairs practically out of question. . . .There is neither time nor opportunity for intimate acquaintance. Instead we notice a trait which marks a well-known type, and fill in the rest of the picture by means of the stereotypes we carry about in our heads. He is an agitator. . . . He is an intellectual. He is a plutocrat. He is a foreigner. He is a ‘South European.’ He is from Back Bay. He is a Harvard man. How different from the statement: He is a Yale man. He is a regular fellow. He is a West Pointer. He is an old army sergeant. He is a Greenwich Villager: what don’t we know about him then, and about her? He is an international banker. He is from Main Street" (Lippmann 1922, Chap. 6, Sec.3).

Modern pollsters tend to agree with this classical scholar of public opinion that many of their answers from the public are stereotyped. This reduces the value of public opinion compared to expert opinion.

·        The early students of public opinion such as Lowell and Lippmann conceived of public opinions as attributes of lasting and functioning networks. They assumed that a public had such a density that the participants could talk and argue about a common issue so that everyone's view became known and influenced by everyone else's view. The result of this process they called public opinion, not necessarily unanimous, but one that emerges after all sides had been heard and considered. Today public opinion is measured by interview surveys of the public. The results of the encounters between a random-like selection of respondents and a questionnaire writer (sometimes mediated by an interviewer) are what the pollsters call public opinions. This is the major difference between polled opinion and public opinion in democratic theory.

The pollsters cannot claim that the opinions they record are properties of lasting networks, and/or, that the opinions they report have emerged in discussions in networks. What they may safely say is that the opinions they recorded are expressed in temporary encounters. This encounter may be a person-to-person meeting, or phone conversation between an interviewer and a respondent, or it may be an indirect encounter between a questionnaire writer and a respondent mediated by post or email. As we have noted, there is much debate and methodological research among survey researchers on the merits of these various forms of encounters for data collection. What concerns us now is something different: any such encounters are smaller and more fleeting forms of interaction than those in networks. This is also true about encounters in the form of pre-arranged focus groups. They provide for more ordinary give-and-take conversations than the interviews, but they do not recur on subsequent occasions as do interactions in most real-life networks. James Fishkin’s (1991, 1995) effort to renew opinion research by gathering respondents for a weekend and turn them into audiences for political education, and subject them to repeated interviews, is not a full bridge to classical public opinion. After the event, the participants return home, never to meet again. In all, the opinions recorded by pollsters are not anchored as firmly in the social structure as are the opinions in networks, i.e. what the classical theorists called public opinions.

Theory of opinion needs to be revised so that its definitions come closer to the actual practice of polling. And the practice of polling needs to be revised to meet the requirement of a viable democratic theory.

Stages of Opinion Maturity

Opinion polls can easily obtain answers that show that people think that automobiles are an environmental hazard, and that people support measures to bring down their environmentally dangerous exhausts. But all hell breaks loose, not least in environmentally friendly and sparsely populated areas of great distances such as Alaska and northern Norway, if politicians use such answers as justification for hiking up the tax on gas. A poll can only serve as a guide for politicians if they also investigate how people regard the consequences of their opinions. Most everybody is “against war” when a pollster asks, but few have considered the consequences for the world, their nation, and themselves of not going to war against an enemy like Adolf Hitler.

Daniel Yankelovich (1991) created a bridge – the best we have so far – between polled opinion and classical opinion. He separates immature opinions from mature ones. He discovered that practically every mature public opinion has passed seven different stages. When an opinion has passed through all of them, politicians, business strategists, editorial writers, PR-consultants, and leaders of voluntary associations can trust them.

After a referendum in 1994 Sweden joined the European Union. Sweden, of course, is a part of the geographical and historic Europe, but the country had rejected an earlier opening to join the European Union when neighboring Denmark became a member. Let us take the opinions about joining as an illustration of the Yankelovich stages that public opinion passes through.

  1. Awareness of an issue. As part of the general internationalizations of the late twentieth century, many Swedes realized that they had become increasingly dependent on other parts of the world, and particularly on those countries already in the European Union
  2. Sense that the issue soon must be decided. Openings to join the European Union were far between. The country had to decide before the EU-train with new members had left the station. The next chance lay many years in the future.
  3. Hunting for solutions. The country could cope with internationalization by staying outside the European Union but become more European anyway by increasing its already institutionalized cooperation and integration with its Nordic neighbors. Or, the country could cope with the internationalization by intensifying its work in the United Nations, which is more encompassing than the European Union. Or the country could join the Union.
  4. Wishful thinking. At this stage unrealistic arguments abound. Some say that Sweden has a better welfare system than the other countries in Europe so there is no need to join. Others say that by joining Swedes can reshape the European Union so that it becomes more like Sweden.
  5. Working out realistic choices. At this stage most people begin to realize that there are clear advantages and clear disadvantages – economic, political, and cultural – of joining, and they began to assess the pros and the cons
  6. Cerebral solution. On the pro-side a sense now crystallizes that pro-arguments, and persons and institutions that support the pro-side, are the better ones. On the con-side the opposite sense crystallizes. At this stage of cognitive decision-making the polls in Sweden show even sides. But many who are opposed or uncertain tell the pollsters that they, too, think that the country will eventually join the European Union. This fact hints at the outcome of the referendum.
  7. Mature judgment. At this stage the pros are willing to sacrifice, intellectually and emotionally, the amount of sovereignty needed for a future in the European Union. And the anti-EUs are ready, intellectually and emotionally, to give up economic and other advantages of the Union in order for their country to stay outside and be independent. Both sides think their choice is best for future generations. And both can support their views in debates and conversations even when media present news negative to their views.

Public opinion is shaky and unstable unless it reaches the maturity of the seventh stage. In the Swedish referendum fewer pros than contras reached this stage. A political decision to join the European Union was made, but the majority opinion recorded by the referendum was not mature. For half a decade after the referendum, the Swedish public was more negative to the European Union than any other public in a member country. The pro-EU forces won the referendum but lost, at least temporarily, their cause.

The Swedish referendum of 1994 concerned, among other issues, whether Sweden ought to join the EU in order to influence its development. However, the public had few mature opinions regarding the Union in respect to a common currency, defense policy, veto rights in the Council of Ministers, or whether the EU’s future would be a supra-state (federation) or an inter-state organization (confederation). Some of these questions never reached the first stage of awareness, while others got stuck in the search for solutions and wishful thinking (stages 3 and 4). The polls reported in the media are extremely unreliable at these stages.

Yankelovich holds that editorial writers, politicians, and others cannot reflect a serious public opinion until people have understood the consequences of their answers to the pollsters’ questions. The responsibility to document such understandings rests with the pollsters.

Yankelovich also recommends that when the media publish the results of a poll they indicate the stage that the opinion has reached. This may be an unreasonable demand from a journalistic point of view since it complicates the job of reporting. However, there is no space limitation in the fuller documentation of a poll published on the pollster’s web site. At any rate, journalists should recognize that there are immature and mature opinions. Pollsters must be prepared to answer questions from journalists about the level of maturity of an opinion they have released for publication. It may be hard for them to precisely establish all stages of opinion formation. However, clues can easily be given. How many had heard about the issue before the pollster asked? How many had noticed the issue on television or in papers? How many had searched for information about the issue? How many had talked to others about the issue? Did the others in such talks have the same views about the issue as the respondent, or other views? Did the conversation cover the consequences of the various views on the issue, for example, did you discuss what might happen to others if your view prevailed?

On Opinions in Life Spheres

In the early twentieth century Max Weber specified these "life spheres" (Lebensordnungen) for advanced societies: the economic, political, religious, intellectual and, and, at the mico-level, erotic spheres and family life. Each of these is matched by a value sphere (Wertssphär) of particular priorities. In a couple of brilliant lectures on politics and science as professions, he elucidated the competition of the life spheres as a perpetual struggle of demons (Weber 1921, 1922). The spheres tend to become relatively autonomous and develop their own structures with considerable independence from one another (Eigengesetzlichkeit). Let us update these insights.

The bottom of Figure 1 lists some examples of organizations, networks, and media in macro life spheres of society. The area of knowledge, i.e. science in a broad sense (Row 1), is connected with executive, that is instrumental, descriptions, for example, facts and generalizations. Economy and business (Row 2) are connected with executive evaluations, for example, prices and costs. Politics and administration (Row 3) are connected with executive prescriptions, for example, laws and regulations. Art (Row 4) in all its forms deals with descriptive visions that are emotive, that is, expressive. Religions (Row 5) relate to expressive evaluations, for example, ideas about the fundamental value of mankind and the meaning of life. Morality (Row 6) contains expressive prescriptions, ethical rules of conduct. Thus, six communicative acts provide a potential for six fundamental life spheres in human society: economy, polity, science, religion, morality, and art. These form the central divisions of a many-splendored society.

In the life spheres important products that we call their ‘cardinal values’ are created. They are wealth in the economy, order in the body politic, knowledge in science, sacredness in religion, virtue in the realm of morality, and beauty in the sphere of art. Each life sphere also has its special type of freedom: academic freedom, free trade, civil rights, artistic license, freedom of faith, and freedom of conscience. Freedom is implemented in a society, not as an abstract philosophical proclamation; it must be anchored in the routines of the respective life area.

The classical theories of public opinion as well as most modern academic texts on public opinion tend to restrict their vision of opinions to the body politic. Pollsters, however, are happy to record opinions in all major life spheres. They may go into micro-areas such as households and families, and interview about issues such as mate selection and children’s weekly allowance. Pollsters often get involved in market and media research, sometimes also in religious research. They may interview audiences of art museums and sports arenas. They tap the public’s views on moral issues. They may be called upon to evaluate the public’s views about new techniques such as computers and energy systems. My experience as a pollster in tasks such as these taught me hands-on that all issues are not political and all issues do not have political solutions. Polls do not have to be addressed only to the attention of politicians. Many others can benefit.

Of course, political problems have democratic solutions through the use of legislation and new uses of tax revenues. But in the main, the problems of the business community require economic solutions, the problems of knowledge require scientific solutions, spiritual problems require religious solutions, the problems of art and literature aesthetic solutions, and moral problems ethical solutions. Sometimes a political democracy can contribute to these solutions, but as soon as it ventures beyond the political sphere, democracy is seldom able to provide the most essential parts of the solutions.

Democracy is a good organizational form for the sphere of polity. It is the best we have, according to national creeds in the Western countries. Outside of politics, however, the rules that govern democratic decision-making are voluntary and often inappropriate. Whoever desires to substitute conclusions drawn from experiments and research with votes among the institution’s personnel is naturally welcome to do so, but it won’t lead to desired and scientific advances. An employer who wishes to organize his firm so as to let a majority vote among his employees decide the best course of earning money on the global market can naturally do so. But such a firm is in danger of not being able to meet its competition, and on the day bankruptcy is declared there will not be anything left to vote on. Beauty in art can hardly be determined by a majority decision. The same applies to the message of faith and solace offered by religion and to the views on virtue proffered by ethics.

Polls do not have to be addressed to politicians. Many others can benefit. However, the pollsters should not expect that the representatives of the non-political spheres in society take the same kind of interest in polling as do the politicians.


All the concrete parts of a life sphere − a corporation in the economy, a government agency in the polity, a research institute in science, a church in the religious sphere − contain minor elements from other spheres of life (‘institutional embedding’). They cannot function well without them. Nothing in modern society seems to work really well without some of the money from the economy, some regulations from the polity, some commandments of morality, some of the knowledge of science. To achieve an optimum we may also need artistic enhancements and some of the meaningfulness that is offered by religion.

Given this minimum of embedded alien elements from other spheres, each life sphere tends to develop considerable autonomy, what Max Weber called Eigengesetzlichkeit. They have their own rules, own reward systems, own freedoms. They resist interference from the outside.

Niklas Luhmann, the late German sociologist, gave a new and added precision to Weber's concept. Luhmann sees society in a way that is very congenial to opinion research, namely, as an endless exchange of communications and the endless effort to interpret them. His constituent parts of society are, not actions, but communications. He maintains that the life spheres form autopoietic subsystems. Such systems are self-defining, self-evaluating, self-regulating and self-reproducing. Jointly, the systems shape the whole global society. (Luhmann 1984)

An interesting point in Luhmann's theory is that the various subsystems of society certainly can produce disturbances for one another, but that no system, e.g. the economy or the body politic, can successfully intervene in the internal running of another, e.g. science or welfare. Courts belong to the body politic, and legislators can intervene and change the practice of a judge, but only if their proposals respect judicial procedures and traditions. A legislator or an adjuster from a private issurance company cannot dictate to a doctor whether a patient is sick or well, nor how a patient should be diagnosed and treated. At best, they can make arrangements to pay the doctor’s bill without undo arguments. Legislators, employees in insurance companies, and doctors work in different life spheres.

Eigengesetzlichkeit thus sets limits for social engineering. Whenever public opinion supports a measure of social engineering in which political decisions are required to run details inside another sphere of society there is a built in risk of failure. Here is the first instance we record in this paper of an actual perimeter to the power of public opinion, a limitation grounded in science, not in ideology.

On Stratification and Opinion

The prevailing views of stratification in social science are Marxian and Weberian. Class is defined by one's sway in the markets. Marx focused on the sway that gives some people the purchasing power to invest in and own technically efficient means of production, a fact with many ramifications for their future wealth and its accompanying political power. Weber included this, but in his discussions of class he gave more emphasis to people's purchasing ability for their own consumption.

Marx on a grand scale and Weber on a smaller scale assumed that classes are antagonistic. They described class relations as a struggle. Words such as "exploitation" and "oppression" henceforth, have been used by social scientists, sometimes even when class is redefined (as in the present paper) in ways that do not include antagonisms.

Social scientists, particularly survey researchers, usually measure class by recording people's occupations. To be sure, a person's class position is revealed through his sway on the labor market. Some have properties such as farms or shops that provide them with livelihood and job. Others have rents from inherited estates or sinecures from government or churches that make them less dependent on the labor market. The propertyless and unconnected can, in the main, offer only their physical labor. Here, then, is an objective working class. But this occupational category is not necessarily an organized class engaged in a class struggle in Marx' sense.

Weber separated class from power and status. Different organizations, networks, et cetera in the different life spheres have differing visions of what the social order should look like. 'Power' is the likelihood that someone's particular version of order will prevail.

Status to Weber is the honor and esteem in which persons, positions, and lifestyles are held, independently of class or power. He thought that status became particularly important in societies where the acquisition and division of goods was stable. That is, when class is stable, status comes into the foreground.

Both Marx and Weber also applied the sway that defines a person's class, not only to individuals but to organizations, networks, and media, and to anyone with the functions to create, preserve, distribute, and receive a cardinal value. In this way class became a pervasive category in the classical social sciences, related to a very large part of society. No other concepts in social science have claimed as much explanatory power as have class and class conflicts. This hegemony in explanation is not tenable. Corrections are under way. For example, class as the overwhelming determinant of political opinions and political party preferences is under discussion and revision (Clark & Lipset 2001).

But there are other strata and other conflicts. These may be wars between civilizations, nationalities, religions. They can be conflicts between different merchant houses, different bureaucracies, different royal families, between ethnic groups, between invaders and settlers, between persons employed in the public sector and those in the private sector, between men of the cloth and men of science, etc. These conflicts can be as decisive to the course of history as class conflict. Social scientists have therefore, on empirical grounds, tended to reject Marx' idea that the study of class conflicts would have more scientific merit than the study of all other kinds of conflict.

For us, this is a starting point to continue Weber's decomposition of stratification into further distinct dimensions. A fuller multi-dimensional view of stratification in a society sees separate stratifications in the different life spheres and stratifies according to their possession of their cardinal value. Stratification of a total society can be divided into competency, purchasing resources (class), power, taste, piety, moral rectitude, and physical vigor (Figure 3).

Figure 3. Multi-Dimensional Stratification
Life sphere
Cardinal Value







Class (Purchase ability)
















 In feudal society, the main dimension of stratification was power; in industrial society, class emerged as a dominant aspect of stratification. With time, competence in different areas of knowledge has had considerable impact, but usually not as great an impact as political power and economic class. Taste, piety, moral rectitude, and physical vigor also influence a person’s social ranking, but they have nowhere near the same effect in Western societies as have class, political power, and competence.

Those who possess or control the largest shares of a cardinal value are the ‘elite’ of its life sphere. There are seven elites in the many-splendored society, one in each life sphere.

The seven ladders of stratification in Figure 3 cannot easily be reduced to a single one. When the dinner guests from various elites include a professor at a top university, a business tycoon, a parliamentarian, a prima ballerina, a bishop, an international Red Cross official, and a winner of an Olympic medal, then the hostess has an impossible job to place them in rank order at her dining table. To have a good party in this many-splendored society she had better put those people whom she thinks will enjoy each other the most next to one another. She cannot construe a common hierarchy, but she can create exciting interaction in the central zone of her society.

The Central Zone

Every society has a ‘central zone’ (Shils 1982 Chap. 4). This zone is a common term for the organizations, networks, and media in which the political order is welded, to which the wealth of business gravitates, the intellectual centers where research yields new knowledge, the islands of society where art and culture flourish. The central zone is where you find the developers and bearers of the greatest power, wealth, knowledge, culture, and spiritual authority, and where they can meet. In other words, the central zone is where the top strata, the elites, meet and interact. The rest of society is more or less peripheral.

The central zone has nothing to do with geometry, and does not lie in the center of a nation. Nor is it necessarily related to geography. In the United States the business and financial communities are mostly located in New York and Chicago, but the political scene is in Washington. In England, both politics and commerce are centered in the same city, London, while the seat of learning has traditionally been located in Oxford and Cambridge, and Canterbury has been the center for religion. The advances of information technology may, in time, lessen the geographic concentration of central zones. The central zone in a country may be extended by an international layer. The Vatican thus is (or has been) a part of the central zone of Roman Catholic countries.

The beliefs, norms, and values of the central zone were previously limited to neighboring regions. Greater mobility and developments in audiovisual as well as print media have spread them much farther. The penetration of the central zone has been aided by obligatory education and by the enlargement of the market and its effective commercial messages for mass-produced goods and services, including popular culture. New York-Washington-California-Texas is an insidious mental part of the central zone in most parts of the world in a period when the United States is the only superpower.

The historical evidence from Europe's biggest struggle for men's souls, the Protestant Reformation and the Catholic Counter-reformation, was summarized in the saying "cuius regio, eius religio," whoever holds power decides which religion shall prevail. The central zone, the court of the sovereign at that time, set the tone and dogma for the sermons preached in the churches of Europe of the fifteenth and sixteenth centuries.

If the different elites in a central zone have a modicum of other-directedness they are able to respond to one another in a relaxed way. A central zone in which the participants are "sensitized to the expectations and preferences of others" (Riesman 1950, p. 23) is probably more creative and influential than one in which everyone is restricted by the dictates of a gyroscope that records the standards once implanted by their primary groups.

A central zone that sets the tone is a useful image in the study of power. It includes an old-fashioned truth that authority is located, not primarily in persons, but in identifiable central social structures. And the focus on the tone emanating from these central structures, allow us to study the insidious − Foucault (1991) said “secret” − aspect of power. The power over the minds of people is set by symbols from a central zone that seeps into their lives.

Life in the central zone has major consequences for the destiny of a society. It is our hypothesis that the basic priorities in the central zone spreads toward the periphery and in due course tend to become the central priorities also in the periphery. More than anywhere else in their study of society, therefore, social scientists and historians should record the basic priorities of the central zone. Do the elites that make up the central zone favor tradition or modernity? Are their ideas expressions of fundamentalism or pragmatism? Are they materialists or humanists?

Likewise, in the ideal world, opinion researchers ought to record the views of the central zone in a column separate from the national opinions. The nearest we have at present to that is when they report opinion in "the upper middle class." In cities of the United States, studies show that this class tends to originate and implement most of the initiatives of local politics and also most initiatives in other life spheres (Boschken 2002).

The Central Zone and Democracy

Democracy can be viewed as a protest against the domination of the central zone. Democracy demands that the citizens themselves, and only the citizens, be the ruling class. The rules of democracy concern political power: other parts of the central zone are not directly affected by election results. Political scientists today reject all alternatives to democracy. There may be post-modern societies and post-materialist societies and post-nation-state societies, but so far we do not talk about post-democratic societies.

When Edward Shils, the inventor of the concept “central zone,” heard someone say “political scientist” he used to interject “with the scientist understood as in Christian Scientist” (Epstein 1997, p. 2). In the church of democracy, political scientists are priests, per se an honorable calling. However, the theory of the central zone is a science in the true sense of the word; it denotes a confirmed proposition, a law of nature. Here we meet a limitation of the power of opinion. When a force of nature calls and the realities of the central zone collide with the practice of democracy, the latter is modified in order to survive.

And this, in my view, is what has happened. Unanimity about democracy holds firm only so far as an agreement that no one ought to rule against the will of the people. After that, the arguments for democracy diverge.

The oldest line of thinking asserts that in a democracy governments ought to follow the will of a majority of the electorate. This is still stated in many schoolbooks. Of course, Dr. George H Gallup, the most influential of the founding fathers of opinion polling, did not think that the politicians were constitutionally bound to follow public opinion as revealed in his polls. But he felt that they were morally obliged to do so. And he believed that the public's views were loaded with political wisdom, and that a poll was the key to unlock it (Gallup 1944). As a pollster he wanted to ask questions revealing the public's concerns rather than the pollster's concern or the concern of his publishers and clients. He solved this problem in the late 1930s by regularly asking "What is the most important problem facing the country today?" He did not define the issues, his respondents did.

But many political scientists became contemptuous of this view – they called it “Gallup democracy,” a variety of populism. It did not square with the actual process of governing in an actual democracy.

Their initial reframing of the idea of democracy as a realization of the will of the people follows roughly this line of reasoning: Society changes and is modernized; new classes, needs, and interest groups emerge; political activists in these groups form their own political demands; they organize into parties; democratically elected politicians then implement the activists’ ideas, through, for example, reorganizations, subsidies or benefits in legislation. The leader is not supposed to be a commander but rather a chairman who “listens to the party.” When the government has its own agenda, its solutions must first receive consent in the party organization before they can be implemented. Pollsters become engaged in asking about the agendas of the parties, not the agendas of the electorate.

Another way of thinking to accommodate the existence of democracy to ever-present central zones was put forth by the economist Josef Schumpeter (1943). He saw democracy as a competition among elites, with the electorate acting as the jury. The main contents of politics are decided by the ideas of the elites, that is, the ideas of the central zone, not by the jury. This notion fitted particularly well in the European scene in the last century when a landed aristocracy defended their privileges, the clergy guarded moral values, agrarians advocated protectionism, industrial and business elites championed free trade, and strong labor movements demanded welfare rights and a greater share of the national patrimony. When a country becomes democratic, the public becomes the jury in these struggles and decides which elite is to rule. Under this conception of democracy the pollsters began to serve the lobbyists for the various elites.

A final line of thought – state-run democracy – has been developed by several current historians, political scientists, and sociologists. The title of the book Bringing The State Back In reveals the theme (Evans, Rueschemeyer, & Skocpol, 1985). Research on mature democracies shows that the decisive agenda is set by the politicians themselves together with government servants. The state, the staffs of governmental departments and the recruitment of their heads influence politics more than anything else. Political change would thus depend less on changes in popular will, or on how elites think, and more on how the government itself develops, on influences from the party apparatus, and on the situation of government employees. This is where concrete political innovation takes place. Democracy is then redefined to mean an acceptance of this state of affairs, but with the requirement that the results of this innovation be submitted to the electorate in periodic general elections. 

Democracy has in these ways been redefined to meet the realities of a central zone that sets the tone. Here is the second instance we record in this paper of an actual perimeter to the power of public opinion, limitations grounded in science, not in ideology. However, even the minimal forms of democracy have some value. As Karl Popper (1997) has pointed out, they lead to peaceful transfers of power.

The state nowadays generates most political ideas and tests many of them with the help of the polling profession. Ministers and state bureaucracies have begun to sponsor polls and focus groups on a massive scale, particularly in Britain and the United States. Most of this opinion research remains unpublished.

This is a more profound change than the mere funding of the polling industry. It goes into the heart and meaning of the industry’s role in democracy. Pollsters have gone from giving political initiative to the people, to giving it to competing elites, and now to the governments themselves. In this process, the idea that polls exist as an instrument to help turn the public into a ruling class changes into the idea that polls exist to help the central zone and the rulers exercise power and stay in power.

In this emerging situation the integrity of the polling community will be measured, not only by its sampling, interviewing, and question wording. It might also be measured by the questions a pollster does not ask. To stay true to the best polling traditions, a pollster should give full coverage to other topics than those promoted by the government and the lobbies in the central zone.

Societal Creeds

The consensus that develops in the central zone and seeps into the periphery is the 'societal creed.' Gunnar Myrdal made such a creed innermost to An American Dilemma, his study of race in the United States. The American creed is a cluster of ideals centered on liberty, equality of opportunity, and various specific rights such as freedom of speech and private property. They found their first expression among the elite of gentlemen who sought independence of the country from British rule. They were written into The Declaration of Independence, The Preamble to the Constitution, and the Bill of Rights. The political parties could disagree on most anything, but not on the American Creed. The creed spread from the central zone to all parts of the country.

These principles of social ethics have been hammered into easily remembered formulas. All means of intellectual communication are utilized to stamp them into everybody's mind. The schools teach them, the churches preach them. The courts pronounce their judicial decisions in their terms. They permeate editorials with a pattern of idealism so ingrained that the writers could scarcely free themselves from it even if they tried. They have fixed a custom of indulging in high-sounding generalities in all written or spoken addresses to the American public, otherwise so splendidly gifted for the matter-of-fact approach to things and problems. (Myrdal 1944, p. 4)

Immigrants from various backgrounds learned the principles of the creed and accepted them. Even the descendants of the black slaves came to embrace them and talked about America as the land of the free, the land of opportunity, and of the flag as symbolizing the equality of all men. In the United States public support of the Constitution is generally higher than public support for presidents and political parties.

No European country has a societal creed as explicit and extensive as the American Creed. Several European attempts at comprehensive creeds have been discredited by history, such as Mussolini's fascism based on Roman strength, Hitler's Nazism based on the Germanic race, or Lenin's communism based on the working class and the proletariat's right to rule.

Some more lasting creeds in Europe are found in small countries, such as the belief in "neutrality" in Switzerland and Sweden. In these countries any policy that is said to violate "neutrality" is ruled out, often prior to an analysis and a full public discussion. The neutrality creed keeps these countries out of NATO. The creed of neutrality colors opinions beyond the defense of the nation. The spirit of the neutrality creed kept Sweden out of the European Community for many years, and still keeps Switzerland out of the European Union and Sweden out of its common currency.

Creeds tend to be "sacred" and beyond debate. The public opinions they embody are 'untouchable opinions.' Usually pollsters take them for granted and do not study them. Nevertheless they should be encouraged to review them from time to time, to check how exciting or insipid they are.

On Opinions among Makers, Keepers, Brokers, and Takers

There are spurts and suppressions of knowledge, booms and busts in riches, changing balances between order and disorder, shifting styles of art, religion and ethics. The flow of knowledge, riches, orderliness, beauty, sacredness, and virtue through society is the dynamic history of society. To cope with the flow of cardinal values in a scholarly way we shall separate four functions to create, preserve, distribute, and receive them.

Such functions are found in every life sphere. The same person may perform all these functions. But we normally observe a certain structuration of these functions, a division of labor.

In earlier generations most jobs were physical: farmer, fisherman, miller, butcher, blacksmith, carpenter, weaver, launderer, repairman, and so forth. Or, they were jobs with person-to-person service, such as hairdresser, tailor, waiter, nurse, et cetera. In the present generation we must include jobs involving the manipulation of symbols (Reich 1992). Many of the latter jobs are actually of very old standing, but have grown in numbers and importance. Figure 4 gives illustrations based on these fundamental observations:

In this manner we fill in the societal division of labor by examples, as is shown in Figure 4. When this is done we no longer talk about abstract functions of creating, conserving, mediating, and receiving cardinal values but about actual people doing the jobs of a many-splendored society. These people are the ‘Makers,’ ‘Keepers,’ ‘Brokers,’ and ‘Takers.’

Figure 4. The Manning of Functions in the Life Spheres

Life spheres





of cardinal values, i.e. knowledge, riches, order, beauty, sacredness, and virtue
(Illustrative examples)


Researchers Inventors


Teachers Consultants Technocrats




Bankers Insurers

Tradesmen Marketers Advertisers

Consumers Customers


Politicians, Legislators, Civic leaders

Judges Prosecutors Lawyers Police,

Officials Bureaucrats Civic workers

Subjects Citizens


Creative Artists


Actors, Entertainers Exhibitors

Patrons of culture and entertainment



Priests Monks, Nuns

Preachers Zealots Missionaries Miracle workers

Believers Seekers


Creators of high norms, charities etc


Child carers Sick carers Elder carers Moralists

Aspirants to ethical and good living

The cells in this table are “fields” in Pierre Bourdieu's terminology. As distinct from the specific illustrations we have placed in the cells, they are “prepared” for generations of occupants by the actual or potential structuration of society.

Ideological Inclinations

To relate opinions to ideologies is a natural avenue to pursue in opinion research. It is of considerable interest to look at the most congenial ideologies among those who exercise the main functions in a society.

The Makers generally take pride in their distinctiveness. They create new knowledge, new laws, new sources of wealth, new art, new morals, new sacredness. Were everything old to be preserved, or were everyone to be or think alike there would be little or no possibility of creating something new. The congenial ideologies of the Makers are varieties of individualism.

The Keepers are the guardians of the cardinal values, and thus distinguish between knowledge and superstition, wealth and poverty, between that which is legal and that which is illegal, between the beautiful and the ugly, the sacred and the profane, the ethical and unethical. Their most congenial ideologies are based on hierarchy and elitism.

The Brokers emphasize justice and fairness. A teacher should not have favorites, but treat all alike. A storekeeper should not have one price for locals and another for strangers. State functionaries should not discriminate between people, but treat all citizens the same. Theaters, concerts, art exhibits, and museums should be open to all, not just to elites. Most major religions offer salvation to all, not just to a chosen people. The moralist allows no exception to ethical principles. The most congenial ideologies for the Brokers in the various spheres of society are based on universalism and equal opportunity.

The Takers of the cardinal values of knowledge, riches, power, etc. usually insist that everyone partake of them. They follow the norm of a hunt that everyone share in the kill – the person who went along on the hunt as well as the person that brought down the game, those who participated in the hunt as well as those who stayed at home in the village. In the end, everyone should benefit from the hunt. The most congenial ideologies of the Takers are based on a radical egalitarianism that demands an equal outcome, not only an equal start.

All these ideologies are in opposition to the particularism or partiality that prevails when family members or clan members are viewed as more valuable than outsiders. Such partiality persists even in modern societies. Even to a contemporary man it comes natural to say that he values his family more than other men’s families. And contemporary nationalism and chauvinism are replete with elements of partiality: my country, right or wrong! A milder form of partiality can be found in the esprit de corps in certain professions, even those where universalism is dominant, for example, among doctors, lawyers, and teachers.

Regular measurements of shifts in ideologies are an interesting part of the polling endeavor.

The Era of the Takers

As we entered a new century, the 1900s were called the “century of the common man.” Others have called it the “the century of democratic man,” and cite the many entitlements people have won as the principal sign of progress in our time. The advances of political democracy, women’s liberation, the solidarity of the welfare state with the weaker members of society, and aid to poor nations are usually cited as examples to back up this claim. An excellent indicator of the unprecedented position given to the common man is the proliferation of opinion polls.

In our terminology, the Makers in the economy and the Keepers in the body politic, that is, Weber’s capitalists and bureaucrats, dominated the first quarter of the twentieth century, but the Brokers and, even more, the Takers, have increasingly dominated the rest of the century, particularly in Western societies.

A first analysis of the progress of the Takers was presented as early as 1929 by the Spanish philosopher Jose Ortega y Gasset under the title La rebelion de las masas. Since the beginnings of agriculture mankind has known that as ye sow so shall ye reap. The first generations of Europeans who left their farms, built colonial empires, and introduced industrialism were energetic, individualistic spirits who took charge of their own destinies. But the majority of Europeans of the twentieth century turned out differently. They developed into what Ortega called “unqualified masses,” who did not demand much of themselves, but demanded all the more for themselves.

In Ortega y Gasset’s view, these people are ignorant about what the creating, conserving, and mediating institutions have done for their societies. In other words, they usually exist in blissful ignorance of all the efforts, investments, and systems that have given us a life full of comforts and freedom of choice. They seldom recognize that these blessings of civilization are the fruits of enterprising spirits, of their knowledge, techniques, and social innovations, which in turn depend on the energy, foresight, and resoluteness of forerunners. Instead, they demand priority in receiving such blessings, as if they were theirs by right. The unqualified do not necessarily desire to become qualified; they merely want the advantages of the qualified.

The unqualified masses consider that they have a natural right to receive the good things in life. The declaration of the rights of man of the United Nations (which was not in existence when Ortega y Gasset wrote) is a consummate catalogue of rights to the good things in life. But irrespective of whether rights are formalized or not, the unqualified masses insist that their comforts be provided, their wishes and impulses fulfilled.

Ortega y Gasset’s term “unqualified masses” has a negative connotation, and so has his notion that the masses have a psychology that is reminiscent of that of a spoiled child. I prefer to call them “Takers.” They are, objectively speaking, the takers of the cardinal values of society, including virtues and aspirations to ethical and good living. They were much repressed in earlier centuries, their rise to prominence in the twentieth century is easy to defend.

In a period when Takers reach visibility, the Brokers rise in importance, particularly the helping classes, who serve the Takers. At the height of the expansion of the Danish welfare state this phenomenon was analyzed by Jorgen C. Rich (1973) in a book he called – without being facetious – The Ruling Class. The new ruling class consists of state employees in the social sector, higher education and the health services. Like all ruling classes they command a great share of the resources of society. Their power is based, not on ownership, but on their capacity to create a compelling social ideology of radical egalitarianism with roots in a humanitarian ethic of support to "all the weak in society." Elements of their ideology in Denmark included a rejection of manual work, and fear of unemployment, illness, and death. It manifests itself in perfectionism and universal appeal ("we may all be weak one day"), and a social criticism that protects the interests of the helping classes through good salaries, limited work effort, and a prodigious expansion of the public sector.

Figure 5. Polling Support for a Generous Welfare Program among Makers, Keepers, Brokers, and Takers.
Chart of absolute number of interviewees, construed for purpose of illustration. N=1500.

Figure 5 illustrates the theoretically derived support for a generous program of social welfare. Takers are the largest groups in terms of number of persons; together with the Brokers, they gained the most when democracy was introduced. These two groups give such programs the strongest support. 

However, in terms of cardinal values, the other groups – Makers and Keepers – contribute more than do the Takers and Brokers. This discrepancy leads inevitably to conflict and debate on how much weight to give to polls reporting overall majorities. Books like Robert Weissberg’s Polling, Policy and Public Opinion. The Case against Heeding the "Voice of the People" (2002) are bound to appear. Weissberg singles out polls on expanding social and medical welfare to the general population as unusable guides for politicians in the process creating a welfare policy. Much of his critique hits polls reporting opinions as immature as a Christmas wish list. We have seen in a previous section of this paper that advanced modern polling can cope with such problems and report when opinions are mature. More serious is the argument that Weissberg shares with Rich, that an expansion of welfare supported by polls oversteps the borderline where the cost exceeds the benefit to society. The result is both social disparagement and economic exploitation of the rest of the population.

It is in the nature of the case that Takers should love publicly financed welfare, and when asked, they naturally want more of it. The costs of all their combined wishes could force governments to raise taxes and /or borrow money to a level that the Keepers and others think unsafe. In this sense, it is actually not very safe for politicians to heed a solo voice of Takers when it comes to welfare policy.

In my view, any modern society is best described as a confederation of Makers, Keepers, Brokers, and Takers. Each is helpful for the survival and good fortune of the society. Each deserves a fair hearing.

I hope that some day pollsters will learn to sample and report the opinions of the four groups of Makers, Keepers, Brokers, and Takers. The quartet of these four voices is more worth listening to than a unison number of a national majority in a sample dominated by the ubiquitous Takers. A reasonable guide to policy from a poll may be at hand when a particular policy is supported by the majority of all interviewed, as well as majorities within Makers, Keepers, Brokers, and Takers.

Opinion research is a long way from coping with confederative structures found in social theories. The polling industry is actually also a long way from coping with existing conventional confederative structures based on geography. Pollsters are like bulls in china shops when it comes to confederations and federations. No poll in Switzerland gives separate numbers for the cantons in spite of the fact that initiatives and referendums require a dual majority to pass a proposal, one from the voters in the whole country, and one from the number of cantons in which the plebiscite has carried the proposal. (I am told that this is one reason why Neue Züriche Zeitung, the leading Swiss newspaper in the German language, does not publish political polls.) No German poll gives separate numbers from all Länder, and therefore cannot tell the full story of the composition of the second chamber, the Bundesrat. Hardly any coast-to-coast poll in the United States reports separate numbers for each state, in spite of the fact that, regardless of size, each state has two senators, reflected not only in the composition of the Senate but also in the Electoral College that ultimately chooses the President.

Confederations (or federations) are likely future political structures in Europe and other multi-national and multi-cultural areas in the world. The present polling practices of emphasizing total population numbers because it is too expensive to report numbers from confederate members is not an innocent practice. It obstructs rather than encourages a future of a confederate world.



Adler, Renata, 1988. Reckless Disregard, Knopf, New York, NY.

Adorno, Theodore .W., 1991. The Culture Industry, (edited by J. M. Bernstein), Routledge, London.

Bell, Daniel, 1956. "America as a Mass Society: A Critique", Commentary, vol 22, 75-83.

Blumer, Herbert, 1946. "Collective Behavior", in Alfred M. Lee (editor), Principles of Sociology, Barnes & Noble, New York, NY, 166-222.

--"--, 1948. "Public Opinion and Public Opinion Polling", The American Journal of Sociology, vol 71, 535-544.

Bohman, James &, William Rehg (editors), 1997. Deliberative Democracy: Essays on Reason and Politics, MIT Press, MA

Boschken, Herman L., 2002. Social Class, Politics, and Urban Markets: The Makings of Bias in Policy Outcomes, Stanford University Press, PaloAlto, CA.

Cicourel, Aaron V., 1974. Cognitive Sociology: Language and Meaning in Social Interaction, New edition, Free Press, New York, NY.

Clark, Terry Nichols & Seymour Martin Lipset (editors), 2001. The Breakdown of Class Politics. A Debate on Post-Industrial Stratification, Woodrow Wilson Center Press, Washington, DC and Johns Hopkins University Press, Baltimore, MD.

Cooley, Charles H., 1902. Social Organization, Scribner, New York, N.Y.

Downs, A, 1972. “Up and Down with Ecology: The ‘Issue Attention Cycle’”, Public Interest, vol. 28, 38-50

Durkheim, Émile, 1893. De la division du travail social, Alcan, Paris.

Epstein, Joseph, 1997. "My Friend Edward", in Edward Shils, Portraits. A Gallery of Intellectuals, University of Chicago Press, Chicago, IL. . 1-29.

Evans, Peter B., Dietrich Rueschemeyer &, Theda Skocpol (editors) 1985. Bringing the state back in, Cambridge University Press, Cambridge.

Fardon, Richard, 1999. Mary Douglas: An Intellectual Biography, Routledge, London.

Fishkin, James S. 1991. Democracy and Deliberation: New Directions for Democratic Reform, Yale University Press, New Haven, CT.

--"--, 1995. The Voice of the People: Public Opinion and Democracy, Yale University Press, New Haven, CT.

Foucault, Michel, 1991. "Governmentality" in Burchell, Graham, Colin Gordon and Peter Miller (eds.) The Foucault Effect: Studies in Governmentality. Harvester Wheatsheaf, London.

Gallup, George. 1944. A Guide to Public Opinion Polls, Princeton University Press, Princeton, NJ.

Giddens, Anthony, 1990. The Consequences of Modernity, Cambridge University Press, Cambridge.

Goffman, Erving, 1961. Encounters: Two Studies in the Sociology of Interaction, Bobbs-Merill, Indianapolis, IN.

--"--, 1967. Interaction Ritual: Essays on Face-to-face Behavior, Anchor Books, New York, NY.

Gouldner, Alwin W., 1979, The Future of Intellectuals and the Rise of the New Class, Macmillan (Critical Social Studies) London.

Habermas, Jürgen, 1965. Strukturwandel der Öffentlichkeit. Untersuchungen zu einer Kategorie der bürgerlichen Gesellschaft, 2. Auflage, Luchterhand, Berlin & Neuwied.

 --"--, 1981. Theorie des kommunikativen Handelns, Band 2, Surkamp, Frankfurt am Main.

Hyman, Herbert (with a chapter by Charles R. Wright), 1975. The Enduring Effects of Education, University of Chicago Press, Chicago, IL

James, William 1890. The Principles of Psvchology, H. Holt & Co, New York, NY.

Jacobs, Lawrence R. and Robert Y. Shapiro, 2000. Politicians Don’t Pander. Political Manipulation and the Loss of Democratic Responsiveness, University of Chicago Press, Chicago, Ill.

Katz, Elihu & Paul F Lazarsfeldt, 1955. Personal Influece, The Free Press, Glecoe, IL.

Luhmann, Niklas, 1984. Soziale Systeme – Grundriss einer allgemeinen Theorie, Suhrkamp, Frankfurt a. M.

Lippmann, Walter 1922. Public Opinion, Harcourt, Brace and Company, New York, NY.

Lowell, A. Lawrence, 1913. Public Opinion and Popular Government, Longmans Green, New York, NY.

Lyotard, Jean-Francois, 1993. Moralités postmodernes, Galilée, Paris.

McLuhan, Marshal, 1962. The Gutenburg Galaxy: The Making of Typographic Man, University of Toronto Press, Toronto.

Mead, George Herbert, 1934. Mind, Self and Society, University of Chicago Press, Chicago, IL

Merton, Robert K., 1957. Social Theory and Social Structure (Revised and enlarged edition), The Free Press, Glencoe, IL

Myrdal, Gunnar (with the assistance of Richard Sterner and Arnold Rose), 1944. An American Dilemma. The Negro Problem and Modern Democracy, Harpers, New York, NY.

Noelle-Neumann, Elisabeth, 1980. Die Schweigespirale öffentliche Meinung - unsere soziale Haut, Piper, München & Zürich

Noelle-Neumann, Elisabeth & Thomas Petersen, 2000. Alle, nicht jeder: Einführung in die Methoden der Demoskopie, 3rd edition, Springer, Berlin.

Ortega y Gasset, José, 1929. La rebelión de las masas, Revista de Occidente, Madrid.

Park, Robert E. and Ernst W. Burgess, 1924. Introduction to the Science of Sociology, University of Chicago Press, Chicago, IL.

Popper, Karl R., 1997. The Lessons of This Century: with Two Talks on Freedom and the Democratic State, (edited interviews by Giamcarlo Bosetti, translated by Patrick Camiller), Routledge, London.

Rich, Jørgen C., 1973. Den herskende klasse, Borgen, Copenhagen

Reich, Robert B., 1992. The Work of Nations: Preparing Ourselves for 21st Century Capitalism, Vintage Books, New York, NY.

Riesman, David with Nathan Glazer and Reuel Denney 1980. The Lonely Crowd: a Study of the Changing American Character,

Rose, Arnold M. 1949. "Attitude Measurement and the Questionnaire Survey," Scientific Monthly, vol 69, 97-101. Reprinted with amendments in Rose 1954, 293-305.

--"--, 1954. Theory and Method in the Social Sciences, University of Minnesota Press, Minneapolis MN.

de Saussure, Ferdinand. 1916. Cours de linguistique générale, Paris. Translated as Course in General Linguistics, Philosophical Library, New York, NY, 1959

Schumpeter, Joseph A., 1943. Capitalism, Socialism, and Democracy, Allen & Unwin, London.

Shils, Edward, 1982. The Constitution of Society, University of Chicago Press, Chicago, IL.

Smith, Ted J. & J. Michael Hogan, 1987. “Public Opinion and the Panama Canal Treaties of 1977”, Public Opinion Quarterly, vol 51, pp. 5-30.

Sudman, Seymour. & Norman M. Bradburn, (1982) Asking questions. Jossey-Bass Publishers, San Francisco, CA.

Thompson, Michael & Richard Ellis & Aaron Wildavsky, 1990. Cultural Theory, Westview Press, Boulder, CO.

Tourangeau, Roger & Lance J. Rips & Kenneth Rasinski, 2000, The Psychology of Survey Response, Cambidege University Press, Cambridge.

Weber, Max, 1921. Gesammelte Politische Schriften, Duncker & Humboldt, München.

--"--, 1922. Gesammelte Aufsätze zur Wissenschaftslehre, J.C.B. Mohr (Paul Siebeck), Tübingen.

Simmel, Georg, 1908. Soziologie, Untersuchungen über die Formen der Vergesellschafung, Dunker & Humbult, Berlin.

Weissberg, Robert, 2002. Polling, Policy and Public Opinion. The Case Against Heeding the "Voice of the People", Palgrave Macmillan, New York, NY.

Wildavsky, Aaron B., 1971.The Revolt Against the Masses, and Other Essays on Politics and Public Policy, Basic Books, New York, NY.

--"--. 1991. The Rise of Radical Egalitarianism, University Press of America, Lanham, MD.

Yankelovich, Daniel, 1991. Coming to Public Judgment: Making Democracy Work in a Complex World, Syracuse University Press, Syracuse, NY.

Zetterberg, Hans L., 1992. "Medien, Ideologie und Schweigespirale" in Jürgen Wilke (editor), Öffentliche Meinung. Teorien, Metoden, Befunde, Verlag Karl Alber, Freiburg/München:, pp 51-75




Public Opinion.

In the Spirit of Rousseau.

Creating Source Material for Opinion Research by Means of a Questionnaire.

  The Language in Questionnaires.

  The Structure of Questionnaires.

An Encounter Known as The Interview.

On Opinions in Encounters.

Opinion and The Motivation To Look out for Ourselves.

Opinion and The Motivation To Look out for Others.

Explaining Falling Response Rates.

Opinions and Norm Enforcement.

  Moral Panics.

On Opinions in Positions and Roles.



Accounting for Opinion by Background Factors.

On Opinions in Organizations, Networks, and Media.


  The Staffs of Organizations.

  Freedom of Opinion in Organizations.



  Editors and The Journalistic Process.


The Structuration of Issues.

The Interplay of Organizations, Networks, and Media.

  Crossing  Media and Networks

  Crossing Organizatons and Networks

Classical Public Opinion and Modern Polled Opinion.

  Stages of Opinion Maturity.

On Opinions in Life Spheres.


On Stratification and Opinion.

The Central Zone.

The Central Zone and Democracy.

Societal Creeds.

On Opinions among Makers, Keepers, Brokers, and Takers.

Ideological Inclinations.

The Era of the Takers.




[*] The author, Hans L Zetterberg, is a Swedish sociologist, public opinion analyst, and publicist. He has previously taught theoretical and classical sociology at Columbia University in The City of New York and at Ohio State University. He has also been a publisher of scholarly books (Bedminster Press), the chief executive of a major foundation supporting social sciences (The Tri-Centennial Fund of the Bank of Sweden), a long-time pollster and market researcher (Sifo AB), and the editor-in-chief of a large newspaper (Svenska Dagbladet). Together with his wife, Karin Busch Zetterberg, he now runs ValueScope AB, a consultancy on social research. He is a past president of WAPOR. His web archive (where also this paper is found) is at and mail can be addressed to His current project is a book entitled The Many-Splendored Society. This paper draws in part on the book manuscript and is therefore copyrighted as noted here: © Hans L Zetterberg.

Greta Frankel has translated several fragments included in this paper that were originally written in Swedish.