Daniel McDonald. Handbook of Media Studies. Editor: John D H Downing, Denis McQuail, Philip Schlesinger, Ellen Wartella. Sage Publication. 2004.
“We don’t know that what we’re saying is particularly significant, but it is at least true.” ~ Robert K. Merton (1957, describing mass communication research)
Over the past 50 years, a number of authors have attempted to review the literature or offer conceptual schemes for classifying media effects (Hovland, 1954; McLeod & Reeves, 1980; Roberts & Maccoby, 1985; Weiss, 1969). Lazarsfeld (1948a) summarized the problem well:
This dearth of substantial results is due to the difficulties of the field, which become apparent as one realizes what a complexity of problems the simple term effect produces. Mass media can affect knowledge, attitudes, opinions and behavior of individuals. These effects can be immediate or delayed, of short duration or long-lasting. Effects upon individuals might slowly become transformed into institutional changes. They can come about in simple reactions or complicated chains as when institutional changes produced by the media in turn affect individuals. (p. 249)
Along with variations in the conceptualization of the nature of what we mean by effects, studies of media effects may variously consider the mere existence of the media or a particular medium (Centerwall, 1989; McLuhan, 1964; Meyrowitz, 1985), the special characteristics of the media or a medium (McClure & Patterson, 1974; Munsterberg, 1916), the content of media or a medium (Gerbner & Gross, 1976), or a specific factor within certain forms of communication (McLeod, 1995). All of these conceptualizations, as well as many others, have been described as the causal agent in media effects studies. In addition, the notion of causality itself has undergone multiple changes in social scientific philosophy (Owens, 1992; Salmon, 1989). With these limitations in mind, this chapter will provide a brief historical overview of U.S. media effects research in the 20th century.
The Pioneer Phase
Early in the 20th century, the field of social psychology was emerging at the crossroads between sociology and psychology. A small number of researchers and theorists began to explore such ideas as the psychology of being a member of a social group, group dynamics, and the impacts of the psychology of individuals on the organization. The era was also an important one in development of the mass media, for it saw the emergence of motion pictures, radio, and the telephone as mass communication devices. A few pioneering psychologists, sociologists, and social psychologists wrote about the impact that the new media were having on audiences and on society. Their writings and research set the groundwork for much of how we conceptualize media effects today.
Charles Horton Cooley, one of the first social psychologists, was one such theorist. In his extraordinary work, especially Human Nature and the Social Order (Cooley, 1902) and Social Organization (Cooley, 1909), Cooley set for himself the task of explaining the role of communication in society. First and foremost, Cooley saw the new communication media as providing an expansion of what has gone on before. For Cooley, the new media had certain defining characteristics, and those characteristics should indicate the types of effects we can expect.
Cooley suggests that the new media seem to be most clearly making gains over traditional communication in terms of speed of transmission and diffusion through the social classes. He noted two opposing strains as the result. On one hand, the media will encourage individuality by offering ideas that are congenial to a person’s self-interest. On the other hand, the new media break down limits to the spread of ideas and customs, leading to a universal assimilation and sameness. Thus, Cooley sees the media as fostering individuality and conformity at the same time. The solution to this paradox, Cooley says, is that there are two types of individuality—one of isolation, the other of choice. The media should reduce the former but encourage the latter (Cooley, 1909, p. 93).
For Cooley, the new media have the same type of effects as interpersonal communication—and there is little to distinguish media from interpersonal communication. Starting from the premise that communication is that characteristic that makes us most human, Cooley deduces that what must happen, over time, is a gradual weakening of those things that separate us as individuals, communities, races, and nations as we come to understand that those who look, dress, or act differently are still very similar to us.
The negative media effects that Cooley foresees include the necessity of a public with a rather superficial understanding and concern for issues and other people. This superficiality is brought about by the fact that we learn so much about so many different things that we do not have time to think about, or understand, much of anything in detail. Because our time and attention are limited, he reasons, the addition of new issues to learn about or understand will decrease the amount of time and attention previously spent on fewer issues.
About the same time, Hugo Munsterberg, a psychologist at Harvard University, was thinking about effects that might be attributable to displacing real-world interaction by interaction with the symbolic world of motion pictures. Munsterberg wrote The Photoplay: A Psychological Study (Munsterberg, 1916), in which he combined the physiological aspects of perceiving depth and movement with the psychological characteristics of attention, memory, and imagination to examine what happens when someone views silent motion pictures. This effect is based on the idea that the media provide an interaction with symbols or signs rather than with the objects or people those signs or symbols represent. Munsterberg makes the point that, for us to perceive continuous motion and depth from the series of static images in a motion picture, our brain must integrate the information and produce a whole perception.
Munsterberg (1916) draws parallels between what motion pictures do and what our minds do as we make sense of reality. Just as our minds mold the objective world to our own interests through attention, imagination, and emotions, so too our minds bring together the details of the motion picture to form a coherent story. Motion picture techniques, such as flashbacks and cutaways serve to simulate memory, imagination, and attention processes.
For Munsterberg, then, mediated communication has its impact by presenting material that our mind must accept as real or true so that we can understand the communication. The impacts arise from the manipulation of this symbolic reality and confusion of our perception of what is real with our knowledge that the presentation is completely symbolic. For Munsterberg, this confusion of reality and content is a necessary condition for understanding communication and a crucial part of how mediated communication works.
Other early researchers in this time period were similarly concerned about the effects of motion pictures but primarily concentrated on the effects on children (Bartholomew, 1913; Edwards, 1915; Phelan, 1919). A number of social workers and sociologists of the time were concerned about various aspects of the modern city, such as education, recreation, and leisure (Edwards, 1915; Gulick, 1909; Jones, 1922). Growing delinquency, increases in pregnancy rates, and other issues focused researchers’ attention on potential causes and socialization influences, such as the motion picture and its theaters (Edwards, 1915; Gulick, 1909; Phelan, 1919).
Even in these early days of media effects study, direct media effects were typically conceived of as learning effects; other effects were a consequence of what was learned (Edwards, 1915; Phelan, 1919; Wilcox, 1900). Typically, that secondary effect was thought to be a negative effect (e.g., precocious involvement with the opposite sex, delinquency, learning about things that were “adult” in nature), although the idea that movie content could have a positive effect (morally uplifting or educational) was not lost on these researchers.
By assuming that content led to an effect, researchers were able to bypass the difficult process of relating content to effect and so spent much of their effort documenting the frequency of attendance, social aspects of attendance, and parental perceptions of motion picture effects. By the 1920s, there was a solid mass of evidence indicating that millions of children were attending motion pictures frequently—typically more than once per week—but, from a scientific standpoint, little could be concluded about the effects of motion pictures on those children beyond learning.
The Payne Fund Studies
In 1928, to address the situation more completely, the Motion Picture Research Council set out to determine what effect motion pictures had on children. It did so by obtaining a grant from the Payne Fund, an organization with a history of interest in children and media. The grant was used to support the efforts of the Committee on Educational Research of the Payne Fund. Some of the best-known social scientists of the day were represented on the committee, and they devised a series of studies to investigate various aspects of potential effects of motion pictures on children. The results of that series of studies were completed by 1933 and reflected the state of knowledge regarding the effects of mass communication one third of the way through the 20th century.
Most of the Payne Fund researchers took a very practical stance, simply trying to document the effects of a medium that had, within about 30 years, become a major industry in the United States and around the world. The Payne Fund studies found that children and adults do learn from motion pictures. They could often remember what they learned for a long time afterward (weeks or months later) and that what they learned produced emotional responses. The Payne Fund studies documented many effects that would be restudied in research on radio or television decades later, including such phenomena as “sleeper effects,” miscomprehension, imitation of positive and negative media role models, developmental differences in understanding and learning, and certain aspects of “perceived reality” research, such as the confusion of facts about motion picture reality for the real world.
After demonstrating that high levels of movie attendance were associated with declining morals, delinquent behavior, lower intelligence, and a number of other factors, the researchers faced a question: Does extreme movie attendance lead to conduct that harms reputation, or do children of low reputation go frequently to the movies (Charters, 1933)?
The authors of the studies were unable to answer the question after they raised it. Their conclusion is that there is no simple cause-and-effect relationship. They point toward the idea that all of their data support the notion of a reciprocal relationship—movies do have an effect on children, but those children who are most attracted to the worst movies tend to be those with the most problems to begin with. In phrasing that would echo throughout the later history of media effects research, two of the researchers noted that motion picture influence is specific for a given child and a given movie and that the same picture may influence different children in opposite directions (Shuttleworth & May, 1933).
The primary contribution of the Payne Fund studies was to document, as had not been done adequately earlier, that children and adults do learn from media and that what they learn has an impact on how they live their lives. Although the researchers were not able to specify exactly who would learn what or how learning affects behavior, it was clear that a simple process was not a viable explanation.
The Radio Era
By the time the Payne Fund studies were published, a new medium was garnering research attention. Network radio had begun in 1927 and, by 1935, had millions of listeners each evening. Unlike the motion picture audience, for which attendance figures could be calculated directly from sales, the radio audience was invisible. Ratings services had been developed to estimate the size of the audience for particular programs to enable advertisers to get a sense of who was listening (Beville, 1985). In 1933, President Hoover’s commission on social trends compiled a two-volume set of studies devoted to understanding modern life. Recent Social Trends included one chapter devoted to the impact of radio, listing more than 100 effects of radio—as determined by the study’s authors—on American society (President’s Research Committee, 1933).
In 1935, Hadley Cantril and Gordon Allport published The Psychology of Radio, an attempt to clarify some of the descriptions of effects appearing in Recent Social Trends and take stock of what was known about radio, as well as what kinds of effects we might anticipate through radio. The authors noted that the medium’s blending of interpersonal and impersonal characteristics should produce unique social effects. In their final chapter, Cantril and Allport describe the comparison of the degree of social participation afforded by radio and other forms of communication.
Given the thoughtful and careful nature of The Psychology of Radio, it is a bit ironic that Cantril is now better known for his study of an event that occurred in 1938, on the eve of World War II, when it became clear that the media were capable of achieving massive, dramatic effects.
Theater director Orson Welles was hosting a 1-hour weekly radio program in which works of literature were adopted for broadcast. For their Halloween broadcast, Welles and his writers had adapted H. G. Wells’s The War of the Worlds, a science fiction story of an invasion by Martians. Because of a number of difficulties in working the original story (set in England) into a 1-hour U.S. radio broadcast, the setting was changed to New Jersey, and a somewhat unusual device was used to enable transitions between places and to speed up action: Much of the story, especially the opening, was told through radio news bulletins “interrupting” typical radio content.
Although the broadcast worked well for most of the audience, a small percentage of listeners that night had changed stations to the War of the Worlds from a more popular show a few minutes after the program started. As a result, they missed the opening announcements and thought they were listening to dance music—which was suddenly interrupted by news bulletins. In what was to become a dramatic demonstration of the power of radio to achieve effects, panic was induced among thousands of Americans. The actual extent of the panic will never be known, although some estimates run as high as 6 million who may have believed the Martians were invading the Earth. Cantril and a number of other researchers attempted to understand what had caused the panic (Cantril, Gaudet, & Herzog, 1940).
Cantril and his coauthors attempt to describe the individual characteristics that were associated with panicking or not panicking (Cantril et al., 1940). In other words, probably for the first time, researchers tried to sketch out the types of people and the types of conditions that might lead to such panic reactions. These conditions included such characteristics as suggestibility, critical ability, and fatalism. Such a conceptualization of media effects focused attention primarily on the psychological, rather than sociological, aspects. As the United States entered World War II, such psychological aspects dominated social science concerns as they worked to counteract the effects, or presumed effects, of enemy propaganda (Davison, 1983).
World War II and Communication Effects Research
The War of the Worlds phenomenon must have been in the minds of communication researchers as they assessed mass communication’s role in the world war. In the case of propaganda, whether encouraging patriotism among one’s own troops or hatred for the enemy, the effect was achieved primarily through emotional responses—fear, hatred, pride, love, and other affective responses to communication.
In some ways, those same feelings and emotions that had been roused by the “invasion from Mars” needed to be harnessed and used against an earthly enemy. It was clear that, under certain conditions, the media could achieve very powerful effects. The task of these researchers was to determine what those conditions were. Researchers were asked to determine how to construct the most effective propaganda material. Their attempts to determine the “magic bullets” of effectiveness are documented in The American Soldier (Stouffer, Lumsdaine, et al., 1949; Stouffer, Suchman, DeVinney, Star, & Williams, 1949) and Experiments in Mass Communication (Hovland, Lumsdaine, & Sheffield, 1949).
Although decidedly influenced by the War of the Worlds, Hovland and his colleagues clearly had a different research interest than did Cantril. Cantril was investigating the mechanisms associated with the multiple effects and different interpretations of one radio program, whereas Hovland et al. (1949) were attempting to uncover general principles related to the construction of messages (Lumsdaine, 1984).
For Hovland and his colleagues, the audience provided the measure of success of a persuasive argument. Hovland et al.’s (1949) use of controlled field experiments for gauging the effects of media programs on soldiers’ knowledge, opinions, and attitudes set the parameters for investigations of mass and interpersonal persuasion for several decades. Hovland and his colleagues pursued such issues as differences in the channels of communication (e.g., lecture, documentary films, etc.) and were interested in being able to generalize effects across media—to motion pictures, radio, and newspapers (Lumsdaine, 1984).
Following World War II, it became apparent that there were not any easy answers or magic bullets that could be used in persuasive communication. After the war, Hovland directed the Yale University program on attitude change research, which brought the field study and experimental work they produced to major prominence as methods of studying the social psychology of communication and attitude change (Himmelfarb & Eagly, 1974; Lumsdaine, 1984; McGuire, 1996; Rosnow & Robinson, 1967). The Hovland group (Hovland, 1954, 1957; Hovland, Janis, & Kelley, 1953; Hovland et al., 1949) advanced the study of one-sided versus two-sided arguments and source credibility, among other areas.
Paul Lazarsfeld and his colleagues at Columbia University had begun a program of research on the role of mass communication in modern society in the late 1930s (Katz, 1987; Katz & Lazarsfeld, 1955; Lazarsfeld, 1940, 1946; Lazarsfeld, Berelson, & Gaudet, 1944). After several publications related to the use and effects of radio, their interest broadened to the role of mass media in general and included newspapers, books, radio, and motion pictures. Several volumes of a journal, Radio Research (later Communications Research), were published, documenting their studies. In addition, Lazarsfeld focused heavily on the role of media in personal decision making. Their work dealing with the role of the media in political contexts is still heavily cited today.
Lazarsfeld’s methods (primarily survey research) were very different from those of Hovland and his colleagues, and his conclusions were also very different. Hovland’s research investigated the factors responsible for attitude or opinion change and so often pitted one version or one technique against another, whereas Lazarsfeld’s work was set in the real world of elections, fashion and style, and one person’s influence on another.
Lazarsfeld and his colleagues found that mass media were not often associated with simple directional change in attitude or opinion but were apparently often associated with a reinforcing effect (Katz, 1987). In a sense, the findings parallel those of the Payne Fund and other studies of media violence and behavior—that those predisposed toward violent behavior were most likely to consume violent motion picture content. However, in the case of studies of violent content, the finding was generally considered an indicator of either a powerful media effect or evidence of reciprocal causation. In the case of political content, in which the search was for sources of attitude change, the same type of finding was interpreted as either no effect or reinforcement.
The difference in interpretation of the two similar findings might be explained through examining the connotations associated with the studies. On one hand, in studies of media violence, there is a strong negative sanction against the dependent variable, and any contributory increases in violent behavior are considered negative factors. On the other hand, political attitudes, whether based in typically Republican or Democratic Party ideals, are both equally acceptable and valid, so the only effect of consequence is a conversion from undecided to a decision or the very rare occurrence of a complete reversal from a Republican to Democratic affiliation.
A new form of communication effects research was begun in the 1950s as social psychologists interested in communication as an aspect of interpersonal interaction were developing theories of how communication works. Ted Newcomb’s idea of coorientation (Newcomb, 1953, 1961) focused on one person’s orientation toward another person and how communication serves to increase the accuracy of our perceptions. Drawing on ideas from Charles H. Cooley (1902, 1909), Kurt Lewin (1951), George H. Mead (1982), Talcott Parsons (1953), and Fritz Heider (1958), Newcomb developed a model of coorientation that was to have a major influence on the field of communication a decade later and that continued to exert an influence on the field throughout the rest of the 20th century and into the 21st century (Carter, 1965; Kenny & Acitelli, 2001; McLeod & Chaffee, 1973; Shin & Cameron, 2003).
Newcomb’s idea of coorientation was abstract and, although easily lending itself to most communication situations, the practical application of the principles and approach he developed were unclear. In addition, Newcomb suggested that orientations, which were similar to attitudes, were inferred from behavior (Newcomb, 1961, p. 5), contradicting the then-current idea that behavior followed attitudes. Through the 1950s and 1960s, however, a number of mass communication researchers used Newcomb’s notion of coorientation and other similar models in which perceptions of others played a major role in explaining communication behavior and effects of communication (de Sola Pool & Shulman, 1959; Eisenberg, Monge, & Farace, 1984; Suzuki, 1997; White, 1950).
Fragmentation, Confusion, and Deeper Questions
Reconciling the different ideas, approaches, methods, and findings became a major preoccupation of the late 1950s and early 1960s. The establishment of Ph.D.s in communication led to many more studies of media effects during the 1950s. Publication of Wilbur Schramm’s The Process and Effects of Mass Communication in 1954 highlighted a number of these different approaches to media effects.
Schramm’s (1954) reader provided communication students with research summaries and articles by sociologists, psychologists, anthropologists, historians, and a host of other researchers and theorists. That volume and later editions serve as exemplars of Schramm’s idea that communication is “one of the great crossroads where many pass but few tarry” (Schramm, Riesman, & Bauer, 1959, p. 8), but the collections also highlight how different approaches yielded different answers, how little was known about the conditions under which media had effects, or even about the process of communication itself.
Amid these concerns, in 1959, Bernard Berelson pronounced the field of communication research as withering away, noting that “the innovators have left or are leaving the field, and no ideas of comparable scope and generating power are emerging” (p. 4). Schramm et al. (1959) responded to Berelson’s contention with a range of examples and explanations for what Berelson was observing, but it was clear, even in their comments, that the state of communication research was undergoing a transition period in which research would be different.
Other writers also noted the sea change in communication research. Joseph Klapper’s (1960) book, The Effects of Mass Communication, described a shift from “hypodermic effect” models of media effects to what he termed “phenomenistic” models—those that saw the media as influences within a total situation. W. Philips Davison (1959) suggested that the audience was not a passive recipient of communication but an active, selective partner in the process. Raymond Bauer’s “obstinate audience” (Bauer, 1964, 1965; Bauer & Bauer, 1960; Zimmerman & Bauer, 1956) suggested that a transactional model of give-and-take between audience and communicator is needed to understand communication effects.
Growth in the Discipline, Fragmentation in Direction
By the mid-1960s, media effects research was firmly established within various departments and schools of journalism and communication. Partly because these schools were organized in accord with industry divisions (e.g., advertising, public relations, journalism, organizational communication, etc.), media effects research was similarly fragmented into contexts rather than process. Thus, by that time, effects associated with media were being studied as aspects of journalism, advertising, broadcasting, or whatever department or specialty area in which the researcher was teaching. Even as mainstream psychology and sociology took less interest in studying specific mass media effects, the new field of communication research continued to borrow broad theories and hypotheses from those fields as explanations for communication effects.
In the late 1960s, cognitive psychology began to provide raw material for advances in mass communication effects research. Armed with such concepts as salience and pertinence, media effects research began to abandon the question of whether media had effects to attempt to specify the mechanisms by which those effects were achieved. Across the field of communication research, it became clear, as Chaffee (1977) observed, that media effects often occurred within our minds and were not directly observable; more varied and inventive methods were needed to understand the processes by which media have effects.
Research on Violent Content and Behavior
Increasing crime rates, racial unrest, and generational tension of the 1960s led to a diversion of both researcher attention and grant money from the U.S. government to attempt to understand the causes of violent behavior. During the late 1960s and early 1970s, hundreds of studies investigated the link between media content and violent behavior. A group of studies by social psychologist Albert Bandura are among the best known of these.
Bandura’s social learning theory formed the theoretical base for his (and many other) studies. Bandura (1977) advanced the discussion of media effects by specifying conditions under which he would expect people (especially children) to imitate antisocial and prosocial behaviors presented in the media. In social learning theory, those behaviors that are shown performed by attractive people or those behaviors that are shown to be rewarded are more likely to be imitated. Results of these studies were overwhelmingly consistent. Although some of these same principles were described in some of the Payne Fund studies of the 1930s (Charters, 1933) and even earlier (e.g., Phelan, 1919), Bandura’s important contribution was to make these conditions explicit and test them in an experimental setting.
The success of social learning theory in explaining children’s imitation of media role models led to additional investigations of how and when children imitate media characters. Long-term studies (e.g., Huesmann & Eron, 1986) found additional confirmation of an association of media violence with later aggressive and criminal behavior. The psychological notion of “priming” held importance for Berkowitz and his colleagues (Berkowitz & Geen, 1966; Berkowitz & Rawlings, 1963). They demonstrated that watching a film containing violent behavior, such as a fistfight, might arouse a set of responses that are associated in the viewer’s mind with fighting. If the viewing is followed fairly quickly by a real event that is also associated with the same set of responses as those associated with fighting, the content can “prime” that second set of responses so that it may become more likely that the viewer will respond with a behavior associated with the fistfight.
The preponderance of the evidence pointed fairly quickly to the potential of television to contribute to violent behavior, yet not one of the studies could conclusively show that TV or movies caused violent behavior. Researchers could show that, within a laboratory setting, providing children or adults with violent content and the means to take violent action could result in violent behavior. In the general population, through survey and other studies, researchers could show that children and adults who were violent also tended to watch violent media content. Linking the two in a causal chain became a frustrating and difficult process.
The lack of clear demonstration of causal connections did not go unnoticed. Early on, Feschbach (1956) had advanced the catharsis hypothesis, which suggested that violent television viewing could be a means of reducing frustration and tension, similar to the effects the ancient Greeks believed to occur from attending dramatic theater. Although the catharsis hypothesis received somewhat mixed support and continues to live in some current approaches to mood management (e.g., Bryant & Zillmann, 1994), some researchers, such as Sparks (2002), suggest that the evidence supporting the catharsis effect was primarily a result of problems in the method of investigation rather than evidence for the effect.
By the 1980s, television network executives were frustrated by a large number of studies (done with less than scientific accuracy) that blamed television for a range of violent behavior and attitudes. Centerwall (1989), for example, estimated that a large proportion of U.S. homicides were a direct result of television. A panel of academic researchers was commissioned by NBC to investigate the problem in detail, using an impressive data collection effort and advanced statistical procedures (Milavsky, Kessler, Stipp, & Ruben, 1982). The panel of researchers found little to connect television as a causal factor in later violent behavior. Most academic researchers, though, continued to believe that the evidence, although not completely clear, is strong enough to implicate media violence as a contributory condition. Such a conclusion was reached tentatively in the 1972 Surgeon General’s Report (Comstock & Rubinstein, 1972), and more forcefully in the update sponsored by the National Institute for Mental Health (Pearl, Bouthilet, & Lazar, 1982). In addition, meta-analytic reviews of the media-violence connection during the late 1980s and 1990s provided additional clues to the nature of the relationships and the boundary conditions operating in the connection between content and behavior.
Beyond Behavior: Construction of a Social Reality
Although the question of media as an antecedent or cause of violent behavior remained a popular one, the notion of cognitive effects and affective reactions to media were to dominate the last 25 years of media effects research in the 20th century. Among the highest profile and most researched areas was cultivation theory, developed by Gerbner and his colleagues (Gerbner & Gross, 1976; Gerbner, Gross, Signorelli, & Morgan, 1979) suggesting that television “cultivates” an outlook about social reality in the country. They hypothesize that the more one uses television, the more he or she will accept the “TV world” as reflective of reality.
Cultivation theory generated strong debate and numerous studies throughout the late 1970s and 1980s and remained an active area of research in the 1990s. Tests of the theory became complex and difficult, and modifications to the theory were developed as evidence became difficult to sort out (Gerbner, Gross, Morgan, & Signorelli, 1981). Results and conclusions have been criticized on both conceptual and methodological grounds (Doob & MacDonald, 1979; Hawkins & Pingree, 1982; Hirsch, 1981), and a recent meta-analysis by Morgan and Shanahan (1997) found no statistically significant cultivation effect.
Media Effects and Social Issues
The early 1970s saw the emergence of several major strands of research associated with social issues and political knowledge. Although media effects had been an area of attraction for political science since the time of Lasswell (1948) and before (Wilcox, 1900), researchers in the field typically took ideas from political science and tested them as media effects. In the early 1970s, the field of communication began to develop a number of ideas about the way issues and knowledge about issues were communicated through the media that were to have a major impact on the field and allied areas such as political science.
One of these research areas, the knowledge gap hypothesis, developed by Tichenor, Donohue, and Olien (1970), suggested that, as information diffuses into a social system, certain segments of the population learn the information faster than other segments. The result of the differential rate of learning is that gaps between social groups increase, rather than decrease, over time and with more information. The knowledge gap hypothesis became the focus of hundreds of studies from the 1970s through the close of the century. At first content with replication of the initial findings, later studies began to describe contingent conditions and antecedents of the basic effect (Gaziano, 1985). Although the knowledge gap remained an area of study throughout the century, research in the area peaked in the mid-1980s as more and more communication researchers tended toward psychological, rather than sociological, models for research.
Another area of study during the 1970s and 1980s was stimulated by changing news media. In the 1960s, for the first time, television news had overtaken newspapers in survey responses to the question ascertaining which medium people relied on most for news about current affairs and politics. The announcement generated a number of studies of “media reliance” or “media dependency” in which researchers attempted to unravel the consequences of this change in the social system. Although under experimental conditions, some researchers were able to show that television news was related to political cynicism and disaffection (McClure & Patterson, 1974; Patterson & McClure, 1976). The implications drawn for the social system were quite broad and alarming. A number of researchers demonstrated, through survey data, that those who relied primarily on TV news for their information were not as well informed or as trusting of politicians as those who relied on newspapers.
A flurry of research activity generated in the early and mid-1980s demonstrated similar results. However, because the bulk of the studies did not involve data that would enable causal inferences, some researchers questioned the validity of the conclusions drawn about the role of television and suggested that the findings might as easily be attributed to the type of person who seeks news from television rather than radio.
Studies by McLeod and McDonald (1985), Reese and Miller (1981), and a number of others found evidence that people did, indeed, learn from television news. By the late 1980s, the sweeping conclusions of the earlier research had had considerable doubt cast on it. In a situation that nearly mimics exactly Lazarsfeld’s (1946) study of newspaper and radio reliance, the field essentially concluded that Lazarsfeld was correct in his assertion that those who reported relying on electronic media were less interested in the news in general than were those who reported relying on newspapers (Lazarsfeld, 1940, 1946, 1948b).
One area of study that was developed in the late 1960s and early 1970s illustrates in particular the general trend in media effects research during the past quarter century. Although the idea had been suggested earlier (Cohen, 1963; Lippmann, 1922), the agenda-setting hypothesis was most clearly enunciated by McCombs and Shaw (1972). The hypothesis suggested that the media set the agenda for public discussion of social issues by providing clues about which issues were important to think about.
Probably as a function of the heavy influence of cognitive psychology on media effects research in the late 1970s, agenda-setting research was recast from a social system effect; that is, the media “correlated” responses in social surveys (cf. Lasswell, 1948) to an individual psychological effect in which the media were seen as manipulating the salience of issues. Thus, although the theory originally focused on public discussion as being affected by media coverage, tests of the theory focused on media coverage being associated with issue salience or prioritization in people’s minds. As cognitive psychology had greater impacts on media research in the 1980s and 1990s, a number of additional ideas were added to and complemented the agenda-setting framework (McCombs & Shaw, 1993). Agenda setting was transformed from a hypothesis to a research area, incorporating earlier sociological concerns such as news diffusion (Breed, 1960) and gatekeeping (Becker, McCombs, & McLeod, 1975; White, 1950). Attempts were also made to link agenda setting to other theories, such as news framing, media priming, and the spiral of silence.
Research in news framing is concerned with how issues are presented in the news—which details are important, which are trivialized or peripheral. The area is based on early observations by Gitlin (1980) and studied by a number of researchers in communication and political science (Iyengar, 1991; Iyengar & Kinder, 1987). The research suggests that, by emphasizing certain aspects of issues, framing has implications for understanding the ebb and flow of public opinion.
Media priming effects take the notion of priming from cognitive psychology, which suggests that people can be “primed” to use certain stored sets of knowledge in making decisions or evaluations simply by exposing them to an associated stimulus (Higgins & King, 1981). Media priming is somewhat different from the psychological construct of priming in that it suggests that people are primed to make judgments on such things as presidential performance through repeated exposure to news reports on issues (Kinder, 1998a). The departure from the psychological area of priming is clear in that media priming refers to what might be seen as associative learning—people make connections between presidents and issues, for example, because they are taught those connections through news reports. Repeated presentations of the issues result in audience members’ use of those learned links in evaluating the president. Whereas the psychological construct of priming refers to an activation of mental connections between the prime and the object of evaluation (Higgins & King, 1981), media priming refers to the dominance of certain learned connections in making an evaluation (Kinder, 1998a). Although the idea of media priming has generated a large amount of research, it is a relatively new idea. There are conceptual issues surrounding how it works and how it differs from other related media effects perspectives such as agenda setting or other phenomena such as salience, accessibility, activation, and the psychological concept of priming (Domke, Shah, & Wackman, 1998; Edelstein, 1993; Higgins & King, 1981; Kinder, 1998b).
Media Effects and Public Opinion
A second major research area to emerge in the early 1970s and remain a fruitful avenue for research at the close of the century was in the area of public opinion research. Elisabeth Noelle-Neumann’s (1974) idea of the spiral of silence appeared at first to attract little attention beyond that of members of her institute; however, by the early 1980s, it had emerged as one of the most researched theories in the field. The spiral of silence suggested that people tend to remain silent, rather than express opinions, if they perceive that their opinion is losing ground among the population. Mass media effects within the theory play a fairly minor but important role of providing information about the climate of opinion, and the perception of this climate is what is used in determining whether one’s opinion is gaining or losing ground.
Most researchers studying the spiral of silence pick only certain components to study because the theory requires fairly substantial data to test completely. Although a number of early studies called some of the basic premises into question, empirical studies support various aspects of the theory (Glynn & McLeod, 1984). A meta-analysis conducted in the late 1990s (Glynn, Hayes, & Shanahan, 1997) found that the overall premise had small but statistically significant support. In addition, in one of the few studies testing the theory over time, McDonald, Glynn, Kim, and Ostman (2001) used the 1948 election data collected by Berelson, Lazarsfeld, and McPhee (1954) and found support consistent with the premises of the theory.
Media use as Effect
As the first century of empirical mass media effects research in the United States was drawing to a close, research on the internet loomed large in journals and the popular press. A number of these studies are simply replicating older studies done with different media (see Wartella & Reeves, 1985). Some of these attempt to incorporate what has been established about our interactions with media. The best of these merge the interpersonal communication literature with mass media effects literature in a way that may not be too far off from what Cooley (1902, 1909) imagined at the beginning of the century and also incorporating more of the social element that had largely been missing from media effects research since the Payne Fund studies. The worst of them sink to what Chaffee (1979) described as the “synthetic competition”—a pitting of one medium against another to determine which is “best.”
During the 1980s, physiological research methods began to be employed by a number of researchers working on media effects. Although some types of physiological measures were being used as early as the Payne Fund studies of the 1930s, these types of measures were seldom used between that time and the 1980s, when the need to look into the “black box” processes, or at least their outward manifestations, became increasingly important for communication theory. These researchers were most concerned with the areas where media had direct effects: attention, thought, memory, cognitions, arousal, skin conductance, and heart rate. Although the relevance or importance of the variables being studied in this type of research may not be immediately clear, what is clear is that the effect is a result of communication.
In concert with physiological studies that hark back to the early Payne Fund studies, the 1980s began to see affect and affective reaction return as a legitimate area of study in communication research (Dorr, 1981). Research proliferated on frightening mass media, led largely by the work of Joanne Cantor and her colleagues (Cantor, 1994; Cantor & Sparks, 1984; Cantor & Wilson, 1984). Offshoots of this research also examined the role of other people in modifying the mass media experience (Nathanson, 1999).
New Directions
These studies, as well as others investigating how mass media fit into people’s lives, suggest that Bauer’s (1965) ideas of models of the transactional and interactive nature of communication might be on the horizon. As the second century in mass media effects research begins, there are a number of areas in which we may expect media effects research to be conducted over the next few years. Because of their generality, two broad areas may dominate media effects research in all of its contexts: issues of communication and reality and further exploration of the black box.
Issues of communication and reality, distinguishing reality from communication and other similar questions, follow from the earliest investigations of Munsterberg (1916) through the Payne Fund studies (Charters, 1933), Cantril and Allport (1935), Gerbner’s cultivation studies (Gerbner & Gross, 1976; Gerbner et al., 1979), and contemporary issues of “presence” in virtual reality. It is one of the most basic questions in communication research: What is the difference between communication and reality? Although some studies frame the question as if blurring communication and reality is a particular problem, it is clear that blurring is the natural state of things, or else communication would have no relevance to reality. That is, communication is about reality, or it is worthless. The degree to which communication is isomorphic to reality is an extremely difficult question.
From one perspective, no communication is real—it is all communication. In some instances, however, we expect people to act on communication as if they were reacting to a reality—if someone yells “fire,” for example. In other instances, we expect people not to treat the communication as reality—if someone yells “Godzilla,” for example. As the new century of media research begins, we see that it is probably not an issue of deciding whether material depicted in the media is real; it is more likely an issue of deciding which dimensions of reality can be found in any particular communication. The idea of “real” may take on many visual, semantic, and meaning aspects—an affective reality (e.g., illustrating emotions that are consistent with the situation that is being dramatized), a reality of general social principles (e.g., suggesting that people who live fast lives die early), or a reality of visual appeal (e.g., that’s real because it moves like a raptor should move). With computer-generated motion picture locations and characters, the field may find it profitable to study these and many other dimensions involved in how we use, rather than evaluate, communication and reality.
Further Exploration of the Black Box. Understanding how audience members understand reality is a subarea within the broader area of understanding what happens while communicating. Our first hundred years made a number of assumptions about communication; our models divided communication into various components, suggested stages in communication processes, described encoding and decoding of messages, and so forth, but all of these models and theories of communication were based on an assumption that either (a) we would never know the actual processes that occur or (b) the processes could be logically deduced. As we move into the second century of media effects research, we find that the assumptions of (b) are inadequate for a scientific study of media effects, and although (a) may be true in the strictest sense, psychological models and methods for studying information processing have gotten much more sophisticated and moved closer to modeling if not the process then at least the outcome, of thought.