Audience Research in Social Policy
Electronic media play a central role in modern economic and social life. They facilitate the exchange of goods and services, help elect political candidates, and open or close the marketplace of ideas that is so essential to democracies. They may even shape our perceptions of reality. Given the powers commonly attributed to media, it is no surprise that they have been studied by social scientists from a wide variety of disciplines.
Since the earliest days of radio broadcasting in the 1920s, proponents and critics wondered how the medium would affect society By the 1930s, a high-powered government committee on social trends appointed by President Hoover listed more than 150 effects of radio, from homogenizing regional cultures to encouraging morning exercise (Ogburn, 1933). Newly formed networks also began assessments of the radio audience, and academicians, especially from psychology, sociology, marketing, and education, became interested in researching the new medium.
Frank Stanton, a pioneer in communications research who later became president of CBS, published one of the first scholarly studies of the radio audience in 1935. Academics took an immediate interest, especially in the use of media for political purposes in the United States and Germany. They began to examine the use of radio by Franklin Roosevelt, as well as various religious and political demagogues, and they studied the manipulation of motion pictures by Adolf Hitler. In 1935, Hadley Cantril and Gordon Allport of Harvard University published The Psychology of Radio, reporting many of their early findings on topics such as these.
Cantril and Stanton eventually secured a grant from the Rockefeller Foundation to study radio audiences. By the time funds were available, however, Stanton was unable to head the project because he had become director of research at CBS. Instead, they asked social psychologist Paul Lazarsfeld to serve as director. The resulting Princeton Radio Research Project lasted 2 years before moving to Columbia University and becoming the Bureau of Applied Social Research.
The Bureau, often in collaboration with industry, conducted a string of studies that many regard as the real foundation of communications research in the United States. This new field of radio research, especially audience measurement, established Lazarsfeld as one of the founders of the scientific study of communications. In fact, the emergence of ratings research prior to World War II was intertwined with broader developments in the new field of mass communications and the growth of social/behavioral research in general. But electronic media audiences eventually emerged as appropriate subjects for communications research in their own right.
Today, audience data play an important, if underappreciated, role in social policy. They often figure in the work of policymakers, social advocacy groups, and academicians who use them to justify policy positions in the most contentious debates about mass media and society. Ratings data are a particularly appealing tool because they are the heart of the media’s power. Ask yourself: Why do electronic media have economic value? Because they attract an audience. Why do news and entertainment programs have any social impact? Because they attract an audience. Why do political ads influence elections? Because they attract an audience. Even though ratings information alone cannot reveal the effects of media on society, they can frequently index the potential. As we have emphasized throughout this book, ratings are the currency with which media operate, and because of their wide and continuous availability, they are useful for analyzing communications law and social policy.
Among the individuals and institutions most likely to use audience data are (a) the federal government, (b) industry, and (c) the public. The dynamic interactions among these interested parties can determine the course of public policy.
By the mid-1920s, it became apparent that broadcasting could not be left to operate as an unregulated marketplace. Demand for broadcast frequencies far exceeded the supply. The U.S. Congress addressed the problem by creating the Federal Radio Commission (FRC), which was replaced by the Federal Communication Commission (FCC) in 1934. The Commission was tasked with licensing stations to ensure that broadcasters operated in the public interest. Although definitions of “public interest” have changed over the years, three interrelated objectives have endured. First, commissioners have tried to limit certain undesirable social effects, especially among children. Second, they have tried to promote greater diversity in media content by taking steps like structuring markets in a way that makes them more responsive to audience demand. And third, like many other regulators, they have passed rules to ensure the overall economic health of the industry they regulate.
The FCC, however, does not have a free hand to implement whatever policies it chooses. Other institutions within the federal government are often involved. The president, the courts, and especially the Congress can and do make their wills known from time to time. In 1996, for example, Congress passed what many call the most sweeping legislation covering communications services since the original Communications Act of 1934. The Telecommunications Act of 1996 affected many areas of the business, from ownership to content. It required the FCC to review long-standing rules and revise anything that had become outdated.
Other independent agencies, like the Federal Trade Commission (FTC) or the Copyright Royalty Board, also deal with matters of communications policy. Some executive branch offices, such as the Department of Commerce or the Department of Health and Human Services, may also enter the picture. In the early 1970s, for example, the surgeon general oversaw a massive study of the impact of television violence. Studies like this often influence legislators to vote for (or against) particular bills.
The United States government is not alone in using audience metrics to address questions of public policy. Since the 1980s, when reliable peoplemeter systems were put in place, audience data and patterns of media use have affected how European governments address various regulatory issues (e.g., Helberger, 2011; Just, 2009). For example, in 2012, the British Secretary of State for Culture, Olympics, Media and Sports commissioned a report on “Measuring Media Plurality.” Plurality is analogous to the U.S. concern about diversity (e.g., Napoli, 2007). The report concluded that three categories of metrics were relevant to assessing plurality, “but the consumption metrics, especially reach, share and multi-sourcing, are the most important” (Ofcam, 2012, p. 1).
Many nongovernmental interests also weigh heavily in the policymaking process. The organizations with the most direct interest in communications policy are the media themselves. To advance their interests, some companies, such as large media conglomerates, might act as their own representatives in Washington. Often, however, D.C.-based trade associations do this for them. The most significant trade association for broadcasters has been the National Association of Broadcasters (NAB). The NAB serves commercial broadcasters by lobbying Congress and the FCC, testifying before congressional committees, filing briefs in the relevant judicial proceedings, and participating in rulemakings and inquiries at government agencies. In many of these activities, the NAB submits research that bears on the issue at hand. In fact, the association has a special department of research and information that performs policy studies using ratings data. Additionally, two important committees, the Committee on Local Television Audience Measurement and the Committee on Local Radio Audience Measurement, monitor the quality of audience research on behalf of the industry.
The National Cable & Telecommunications Association (NCTA) represents the cable industry. The NCTA engages in the same sorts of activities as the NAB, including policy research. And the Interactive Advertising Bureau (IAB) represents major media and technology companies in the online advertising market. Other trade associations, like the Motion Picture Association of America (MPAA), will occasionally use ratings data as well.
Although government and industry have a major influence on the formation of public policy, they do not completely control it. The public itself enters the process in a number ways. Most directly, we elect government representatives. Occasionally, one of these officials takes the lead on a matter of communications policy, inviting us to either support or reject that position. More highly organized public participation comes in the form of public interest groups. Some, like Action for Children’s Television (ACT), were formed specifically to affect policymaking. Until its closing in the mid-1990s, ACT drew the attention of Congress and the FCC to problems of consumerism in children’s television. Other organizations, like the Parent/Teacher Association (PTA) or the American Medical Association (AMA), express occasional interest in the social control of media even though regulation is not their central focus.
We should also note that the academic community contributes to policymaking. Professors with an interest in broadcasting and the electronic media have been attracted to policy questions dealing with the media’s social and economic impact. They can affect policy in several ways. Most notably, they publish research relevant to questions of public policy. Because they are often viewed as experts, and relatively objective, university-based researchers may carry some special influence with the government. Academicians may also work as consultants for other participants in the policymaking process, exercising direct influence in a less public, and usually less objective, manner. Finally, of course, they can indirectly affect policy through their students. Many professionals who are involved in determining communications policy today would credit certain professors with influencing their views on matters of law, regulation, and social responsibility.
The range of applications for audience ratings in the formation of social policy is reasonably broad. Ratings stand apart from other types of social scientific data because they can be interpreted, or “read,” in so many different ways. Nevertheless, ratings are generally used in policymaking to answer one of three broad questions. The questions correspond to the long-term concerns of the FCC that we mentioned earlier: (a) limiting the undesirable effects of media, (b) promoting more diverse and responsive programming, and (c) tending to the economic condition of its client industries.
What Do the Media Do to People?
Policymakers must answer this “effects question” if they hope to limit undesirable effects. As a policy concern, the question certainly predates broadcasting. In the early 1930s, sociologists tried to determine the impact that movies had on young people. Later in the decade, psychologists studied the effects of wartime propaganda, and marketing researchers questioned the effect of press coverage on voter behavior. More recent studies focus on issues such as television’s ability to distort perceptions of social reality and the role video games play in promoting violence.
Central to these, and all other effects questions, is the cause-and-effect relationship. In its general form, the question is, “Does exposure to the media (cause), make other things happen (effect)?” This is an extremely difficult question for social scientists to answer. An important starting place, however, is knowledge of what people listen to or watch. This is because, by definition, any direct media effect must begin with audience exposure to media messages.
Although an encounter with the media may not determine a particular outcome (again, the effect), hearing or seeing a message does define a certain potential. If many people use a medium, or see an item of content, the potential for effects is great. Advertisers have long realized this fact, and so have paid dearly for access to audiences. The value of this potential is also obvious in the frequently recurring debate over free airtime for political candidates. The opportunity to reach the electronic media audience is perceived by many as a kind of “right” that candidates should enjoy. They assume that citizens’ voting behavior will be influenced by exposure to campaign messages. Conversely, if no one is exposed to a message, its impact is never felt.
Academics, too, have recognized that exposure is the wellspring of media effects (e.g., Bryant & Oliver, 2009). One of the most outspoken scholars has been George Gerbner, a proponent of cultivation analysis. Gerbner argued that television content is so uniform, and people are so unselective, that researchers need only compare heavy and light television viewers to determine the medium’s social impact. Usually, these arguments are buttressed with data on the sheer amounts of time Americans spend watching (Gerbner, Gross, Morgan, Signorielli, & Shanahan, 2002).
Other academic researchers posit a greater degree of audience selectivity. In fact, studies of selective exposure, which we reviewed in chapter 4, have an important place in the history of media research. Although varied in their origins, these studies assume that audience members are capable of discernment and use it in their consumption of media content. Effects depend on the kinds of content they choose. In the surgeon general’s report on television violence, for example, Israel and Robinson (1972) used viewing diaries to assess how much violence was consumed by various segments of the population. They assumed that heavy viewers of violence-laden programming would be more likely to show its ill effects. Whether one considers specific content or, like Gerbner, television viewing in general, exposure clearly sets the stage for subsequent media effects.
Government regulators also have used audience information to assess the media’s potential to create socially undesirable effects. The FCC uses them, for example, to guide its effort to limit indecent language in broadcasting. Although the Commission might tolerate certain excesses when adults are listening, the presence of young children creates a problem. Some policymakers have tried to channel offensive language away from time periods when children are likely to be in the audience. They use ratings to identify those time periods. Hence, the detrimental effects that might result from exposure to indecent content are, at least, limited by the size of the child audience.
Limiting the potentially harmful effects of advertising on children is a concern for regulators worldwide. Rules that reflect this concern often use audience research estimates as a benchmark for policy. In the United States, advertisements for alcohol on television are restricted to programs with an audience profile made up of at least 70 percent viewers aged 21 and older. Additionally, these advertisements cannot run before 9 P.M. This latter restriction is a surrogate for audience demographics—many countries use time periods to restrict the advertising of tobacco, alcohol, and pharmaceuticals (ZenithOptimedia, 2011).
Based on its long history of encouraging localism in broadcasting, the FCC has also expressed concern about the audience for local news and public affairs programming. Their effort has been motivated, at least in part, by a desire to keep people informed about issues of public importance. They were convinced, for example, that cable television would pose a threat to localism by diverting audiences from broadcast stations. In a 1979 report on the relationship between cable and broadcasting, the Commission noted:
Television may have an important effect in shaping the attitudes and values of citizens, in making the electorate more informed and responsible, and in contributing to greater understanding and respect among different racial and ethnic groups…. Historically, the FCC has encouraged particular types of programming—local news, public affairs, instructional programs—on these grounds. To the extent that a change in broadcast-cable policy would dramatically change the amount by which these programs are not only broadcast but also viewed, these issues could be an important component of a policy debate. (Federal Communications Commission, 1979, p. 639)
In some ways, this line of reasoning is a complement to the logic supporting the indecency rules. Instead of people being harmed by exposure to something that is bad for them, the FCC expressed concern that undesirable social consequences might flow from people not watching something that is good for them. Audience measurement can be a significant factor in these evaluations. In 2011, for example, the FCC commissioned a white paper about the consumption of local news on the Internet. Based on a quantitative analysis of comScore data, the study found that online news accounted for a very small percentage of Web traffic (Hindman, 2011).
Today, with the expansion of media options, understanding what people consume or manage to avoid is increasingly important. In 1980, the average household could view nine television stations. Thirty years later, a typical household could get 120 channels of television, not to mention everything available on DVDs, satellite radio, and the Internet. In fact, in 2010 Nielsen stopped reporting the number of “channels” because the digital, multichannel media environment means the term is no longer consistent across households. This abundance has caused some social critics to worry that people will choose to hear only like-minded speech and retreat into the media equivalent of “gated communities” (Sunstein, 2001, 2009; Turow, 1997). If people do, in fact, hear only what is agreeable and never encounter dissident voices, it could have undesirable social consequences. Tracking audience behavior offers a useful way to monitor these developments (e.g., LaCour, 2012; Webster, 2005; Webster & Ksiazek, 2012).
Another important goal of communications policy has been to provide the public with diverse media content. This objective seems very much in keeping with our First Amendment ideals and the benefits that are thought to result from a free marketplace of ideas. But how does one achieve diversity? Although policymakers have different opinions on that subject, the most popular solution has been to structure media industries so that a large number of firms can compete for the attention of the audience. In theory, competitors respond to audience demand by catering to likes and dislikes as expressed in program choices. The more competitors, the more likely it is that various niches will be served. Under this system, ratings can be thought of as a kind of feedback mechanism. Arthur Nielsen Jr. (1988) has described the link between ratings and preferences as follows:
Since what the broadcaster has to sell is an audience to advertisers, it follows that in order to attract viewers, the broadcaster must cater to the public tastes and preferences. Ratings reveal these preferences. (p. 62)
Many commentators find the industry’s argument that they only give the people what they want to be self-serving and deceptive. Most media, they point out, respond to the demands of advertisers, not audience members. Consequently, audiences that are less valuable to advertisers are underserved. Additionally, the advertiser-supported media system requires no measure of the extent to which viewers actually like a particular program, only that they have elected to use it. As we explained in chapter 4, a number of other factors complicate the link between preference and choice. Nevertheless, a considerable body of theory in both psychology and economics assumes that choice is a function of preference, and this provides adequate justification for the use of ratings in policymaking.
The most relevant of these theories has been developed in the study of welfare economics, a branch of the discipline that is concerned with how we can maximize the welfare, or overall well-being, of society. Like other branches of economics, it assumes that people are rational beings who will attempt to satisfy their preferences when choosing goods and services—at least insofar as their pocketbooks allow. Economists refer to this notion as the “theory of revealed preference” (e.g., Varian, 2006). Indeed, they make a case that deducing preferences from behavior may be superior to direct questions about a person’s likes and dislikes. Because advertiser support imposes no direct costs on viewers (i.e., they do not pay a per-program fee), viewer preferences can be freely expressed in program choices. These concepts, and their consequences for how public policy might maximize viewer satisfaction, are fully discussed in Owen and Wildman (1992).
Welfare economists, therefore, have used ratings data to address questions of communications policy. One category of FCC rules that has received intense scrutiny is media ownership. The commission has historically sought to limit certain classes of media (e.g., local newspapers and radio) from owning local television stations, based on the assumption that different owners will contribute different viewpoints to the marketplace of ideas. Unfortunately, existing media may be more adept than newcomers at offering local programming that appeals to viewers. Parkman has, consequently, argued the following:
If these classes of owners produce more popular programming than other classes of owners, the reduction in popular programming should be taken into consideration as cost of the diversification policy. To determine if certain newsgathering organizations are more successful than others in attracting viewers, we can look at the end result that these organizations produce as judged by the viewers, i.e., the ratings. (1982, pp. 289–290)
After analyzing the ratings of local television news programs, Parkman concluded that the commission’s policy imposed “costs on individual viewers by forcing them to choose programs considered by them as less desirable” (p. 295).
The FCC itself has relied on ratings as a kind of “revealed preference.” One notable example is the commission’s designation of stations as significantly viewed. This concept was introduced into FCC rules in the early 1970s and was used to determine the popularity of a signal in a given geographical area. Its definition has affected many areas of regulation, such as must-carry, syndicated exclusivity, effective competition, and compulsory copyright. Although the definition changed somewhat over the years, a station was deemed significantly viewed in a market if it achieved a weekly 2 percent share of audience, and 5 percent weekly circulation in non-cable homes. Estimates like these are made from diary data in most markets, which means they are subject to the kinds of errors we describe in chapter 3. Problems occur when regulators ignore the error estimates and treat these numbers as reliable.
Assessing “what people want” begs the question “Which people are we measuring.” In 2012 Arbitron faced legal action in California because of its sampling procedures with the PPM. The company was accused of undercounting African American and Latino audiences. The complaint claimed that when the PPM technology was introduced, “radio stations that serve African-American and Hispanic audiences lost a disproportionately large segment of rated radio listenership, resulting in a decline in advertising revenue to those radio stations, layoffs, and potential bankruptcy” (Complaint for Equitable Relief and Civil Penalties, 2012, p. 1–2). Arbitron had to reevaluate its sampling procedures and assure California’s attorney general that minority audiences would be fairly represented.
What Are the Economic Implications of Various Policies?
A number of government laws and regulations affect the financial condition of the media and related industries. Because these policies have an impact on the “bread-and butter”—or in some cases the Mercedes and BMWs—of those businesses, they attract the attention of many participants in the policymaking process. Even the FCC, which in recent years has favored increased competition in the media, must remain alert to the economic consequences of various policies. After all, the commission is responsible for seeing to it that broadcasting serves the public interest. If broadcasters are driven out of business by some ill-conceived government policy, the commission’s mandate might be compromised.
Financial statements that describe the media’s revenues, expenses, and profitability are one obvious source of information on the economic condition of the industry. But, for a number of reasons, these data are not always used. For one thing, the commission stopped collecting financial statements from broadcasters many years ago, so the data are not readily available. For another, the harm might be too far advanced to correct economic injury by the time it shows up on company ledgers. A common alternative to a dollars-and-cents measure of economic impact is to use audience ratings. Because ratings measure the commodity that the media sell, policies that adversely affect a station’s audience will damage its economic health. Even though ratings and revenues are not perfectly correlated, evidence of lost audiences is often taken as an indication of lost revenues.
Ratings information has been used in several studies to demonstrate audience diversion from established media. Such analyses are frequently used in skirmishes between broadcasters and the cable industry. Early on in the development of cable, broadcasters claimed economic injury to encourage policies that would restrict cable’s growth. They argued that allowing cable to enter a market would threaten the survival of stations by siphoning off valuable audiences. In 1970, Rolla Park at the Rand Corporation assessed this threat through an analysis of local market ratings data. The results of this study helped shape the FCC’s 1972 rules on cable television.
Again in the late 1970s, the commission considered the economic relationship between cable and broadcasting. And again, Park (1979) and a number of interested parties assessed the state of audience diversion through rather sophisticated analyses of audience ratings information. The FCC made extensive references to these studies in its final report.
The Commission also encountered claims of audience diversion in the context of its rules on syndicated exclusivity. First adopted in the early 1970s, Syndex rules ensured that broadcasters who bought exclusive rights to syndicated programming would not have to compete with cable systems that imported the same program on a distant signal. The imported signal, it was argued, would divert audiences that rightly belonged to the local station. In subsequent debates over the rule, the parties at interest (e.g., NAB, NCTA) submitted analyses of ratings data to show that audience losses did or did not occur in the absence of the rule. Although the rule was dropped for some time, the FCC reimposed it, reasoning that “the ability to limit diversion means broadcasters will be able to attract larger audiences, making them more attractive to advertisers, thereby enabling them to obtain more and better programming for their viewers” (Broadcasting, 1988, p. 58).
More recently, ratings data have been used in policy research to analyze television ownership restrictions. The Telecommunications Act of 1996 granted owners of UHF stations special consideration with regard to total audience reach. As stipulated in the Act, the combined audience reach of all stations owned by a single person or entity cannot exceed 35 percent of the total United States. However, because UHF stations have traditionally operated at a disadvantage compared with VHF stations, lawmakers allowed UHF owners to “discount” their reach estimates. They can count just half of the total station coverage towards the 35 percent total. Thus, a group owner of all UHF stations might have potential coverage of more than 50 percent of all television households, but under the FCC rules that would count as only 25 percent. To counter a challenge to this arrangement, the NAB prepared a report showing that UHF stations consistently drew smaller audiences because they operate on the UHF band. After accounting for other factors that could cause lower ratings, the NAB found that channel assignment was related to lower ratings. For example, UHF stations affiliated with Fox earned an average of 1 rating point lower than their VHF counterparts, and NBC affiliates demonstrated a difference of 3.6 ratings points between UHF and VHF stations (Everett, 1998).
Ratings data also have had a substantial impact on the distribution of fees from the compulsory copyright license. These fees are paid by cable and satellite systems for the right to carry broadcast signals. Those with a claim on the resulting revenues are copyright holders including program suppliers, commercial broadcasters, public broadcasters, and Canadian broadcasters. It is logical that audience shares would figure in the computation of awards. After all, the economic value of a program or program service rests largely on its ability to attract an audience.
The uses of audience information in legal or regulatory proceedings are considerable. Despite these, and many other applications of the data, it appears to us that social scientists have only scratched the surface of the analytical possibilities. For the most part, these uses of the ratings have dealt with gross measures of audience size. Perhaps that should not be surprising, because such estimates are the most readily available. Indeed, that is what the ratings are. Using ratings to track individuals over time, engaging in what we call cumulative analyses, would seem a logical next step for social scientific inquiry.
Take, for example, the question of media effects. Although the size of the audience exposed to a message may suggest something about its potential effect, so too does the regularity of exposure. Advertisers recognize this concept as frequency—the average number of times audience members see or hear a message. Effects researchers might similarly ask how often people see or hear a particular kind of programming. For example, do all children see about the same amount of violence on television, or do some consume more? Is some segment of the child audience violence junkies? If so, who are those children? Do they come from poor or affluent families? Do they watch alone or with others? The answer to such questions, all of which can be gleaned from audience data, might contribute much to our understanding of the impact of televised violence. Similar questions could be asked about the audience for news and information (e.g., Kim & Webster, 2012; Ksiazek, Malthouse, & Webster, 2010; Prior, 2007).
Studies of audience duplication might reveal more about people’s preferences for programming as well. Does a particular program have a small-but-loyal following, or is it just small? Programmers and marketing researchers have long recognized a certain feature of audience duplication called channel loyalty. Religious, Spanish language, and at least some news channels are among the kinds of programming that seem to attract small-but-loyal audiences. Does this intensity of use suggest something about how the audience values a service above and beyond the number that uses it at any point in time?
Factors other than size and composition also affect the economic value of an audience. Advertisers may specify reach and frequency objectives in their media plans. Those who seek a high frequency of exposure might be willing to pay a premium for that small-but-loyal audience. In a similar vein, channel loyalty and inheritance effects undoubtedly contribute to the audience of a syndicated program. If a station adds value to the program by delivering an audience predisposed to watch, then perhaps the station should have a greater share of credit for a program’s success.
Even media critics who distrust social scientific methods might learn more about audience experience with media through inventive uses of ratings data. For instance, analysts of popular culture have become increasingly interested in how people make sense of television programming. One insight from this line of research is that viewers experience the medium not as discrete programs, but as strips of textual material called flow texts. The emergence of flow texts could be studied through analogous research on audience flow.
All of these analyses, and many more, could be realized through the application of commercial audience data. Unfortunately, the effective use of such data in the social sciences and related disciplines has been uneven. In part, that is because the proprietary nature of syndicated research makes it too expensive for strictly academic analyses. Some academicians, however, may fail to exploit the data that are available, simply because they do not recognize the analytical possibilities.
Baker, C. E. (2002). Media, markets, and democracy. Cambridge: Cambridge University Press.
Bryant, J., & Oliver, M. B. (Eds.). (2009). Media effects: Advances in theory and research (3rd ed.). New York: Routledge.
Comstock, G., & Scharrer, E. (1999). Television: What’s on, who’s watching, and what it means. San Diego: Academic Press.
Lowery, S., & DeFleur, M. L. (1995). Milestones in mass communication research (3rd ed.). White Plains, NY: Longman.
Napoli, P. M. (2007). Media diversity and localism: Meaning and metrics. Mahwah, NJ: Lawrence Erlbaum.
Owen, B., & Wildman, S. (1992). Video economics. Cambridge, MA: Harvard University Press.
Sunstein, C. R. (2001). Republic.com. Princeton, NJ: Princeton University Press.
Sunstein, C. R. (2009). Going to extremes: How like minds unite and divide. Oxford: Oxford University Press. 2009.
Turow, J. (1997). Breaking up America: Advertisers and the new media world. Chicago: University of Chicago Press.