National Television Violence Study

views updated

NATIONAL TELEVISION VIOLENCE STUDY

Violence on television has been the subject of debate for decades in the United States. It seems as though everyone has an opinion on the topic. Many observers argue that there is an excessive amount of bloodshed on television. In fact, a 1999 national poll by the Pew Research Center found that 70 percent of Americans believe that entertainment programs contain too much violence. Others criticize certain types of portrayals that seem overly graphic or gratuitous. Still others defend the use of violence in the media by pointing to movies such as Schindler's List (1993) and Saving Private Ryan (1998), both of which contain a great deal of physical aggression but have educational value.

Spending an evening with the television remote control can fuel this debate. After the 1999 shooting at Columbine High School in Colorado, Josh Getlin (1999, p. A17), a reporter for the Los Angeles Times, described television in the following way:

Scenes of unspeakable carnage from Columbine High School (click) gave way to images of buildings burning in Belgrade after a NATO attack (click) followed by a hidden-camera video showing a nanny beating a toddler (click), then a Western shoot-'em-up (click) and more scenes from the suburban campus where students were gunned down like targets at a carnival arcade. It was just another night on American television and a disturbing reminder of how deeply ingrained violence is in our culture.

Is the television landscape truly saturated with violence? Does all violence on television look the same? In 1994, researchers at four American universities set out to answer these questions. The result is the National Television Violence Study (NTVS), a comprehensive, scientific analysis of the nature and amount of violence on American television. The researchers monitored more than eight thousand hours of television across a three-year period, from 1994-1995 to 1996-1997. Three annual reports were released as part of the project (National Television Violence Study, 1997, 1998a, 1998b).

The study was a milestone in the history of television research for two reasons. First, it went beyond simple counts of violent behaviors in a program and instead provided an extensive analysis of how violence is portrayed on television. Contextual features such as whether a perpetrator is attractive and whether the violence is rewarded are far more important than the sheer amount of aggression in trying to understand how harmful a portrayal might be for viewers. Second, the study was based on the largest and most representative sample of television programs ever evaluated in a single project.

Background

The NTVS came about during a decade of intense political and public criticism of the media. During the early 1990s, there were numerous bills before Congress that involved some attempt to regulate television content. In a study of newspaper coverage during the time, Cynthia Hoffner (1998) found that 50 percent of the articles about television violence focused on possible solutions to the problem, most notably, government regulation.

In 1993, in the midst of this public scrutiny, Senator Paul Simon (D-IL) and other policymakers called on the entertainment industry to examine more closely the way in which violence is depicted on television. The National Cable Television Association (NCTA), a trade association representing the cable industry, responded to this call by sponsoring a study of television content to be conducted independently of Hollywood. After reviewing proposals from many expert researchers across the country, NCTA commissioned the NTVS, funding the project at a cost of $3.3 million.

The NTVS involved a team of media researchers from four universities. Researchers at the University of California, Santa Barbara, assessed violence in all types of entertainment programming such as drama, comedy, movies, children's shows, and music videos. It is this portion of the study that will be reported here. Researchers at the University of Texas at Austin provided an in-depth analysis of violence in a specific type of programming—reality-based shows, such as real-life police and rescue shows and talk shows. Researchers at the University of Wisconsin, Madison, studied the role of television ratings and advisories, including their effect on the viewing decisions of parents and children. Researchers at the University of North Carolina, Chapel Hill, examined the effectiveness of antiviolence public-service announcements.

In addition to the researchers, NTVS involved an advisory council, whose role was to protect the integrity and independence of the research, especially since the funding was coming from the cable industry itself. The council also provided the researchers with advice and feedback on the design, findings, and implications of the study. The council included representatives from seventeen national organizations concerned with the effect of television on society. Among these were the American Academy of Pediatrics, the American Medical Association, the International Communication Association, and the National Parent Teacher Association. In addition, one-third of the members came from organizations representing the entertainment industry, such as the Producers Guild of America and the Caucus for Producers, Writers, and Directors.

Foundations of the Study

The NTVS is a content analysis that is strongly based on what is known about how media violence affects viewers. The following conclusions served as the four basic "foundations" for the project:

Television violence contributes to harmful effects on viewers.
Three types of effects can occur from viewing television violence: (a) learning aggressive attitudes and behaviors, (b) desensitization to violence, and (c) fear of being victimized by violence.
Not all violence poses the same degree of risk of these harmful effects.
Younger children are a special audience.

These foundations resulted from an extensive review that the researchers conducted (before beginning the project) of all the scientific studies that had examined the effects of television violence.

Foundation 1

The conclusion that television violence contributes to harmful effects on viewers has been reached by virtually every major professional organization that has reviewed the research, including the American Psychological Association (1993), the American Medical Association (1996), and the National Institute of Mental Health (1982).

Foundation 2

Literally hundreds of studies show that television violence can contribute to aggressive behavior in children. Moreover, this effect can persist into adolescence and adulthood. For example, one study by Rowell Huesmann (1986) found that exposure to television violence at the age of eight years helped to predict criminal behavior in a sample of adults. In addition to increased aggression, there are two other types of harmful effects that can occur. Repeated exposure to violence can cause viewers to become more callous, or desensitized, to the harmfulness of violent behavior. Also, long-term exposure to violent portrayals can exaggerate people's fear of being attacked by violence in the real world.

Foundation 3

There are many ways to depict violence on television. For example, a fistfight can last only a few seconds and be shot from a distance, or it can persist for several minutes and feature many closeups of the action. There are also differences in the types of characters who commit violence—some are heroes trying to save lives and others are criminals acting out of greed or anger. There are differences also in the outcomes of violence—some portrayals show the pain and suffering of the victim, whereas others do not. In other words, not all television violence is the same. In fact, the way in which violence is presented helps to determine whether the portrayal poses harm to the viewer. Certain features of violence increase the risk of a harmful effect such as learning aggression or desensitization, whereas others decrease that risk.

NTVS identified eight contextual features that help to predict the likely effect of violence on viewers. First, the nature of the perpetrator must be considered. An attractive perpetrator is a potent role model, especially for children, and can increase the likelihood that viewers will learn aggression from a portrayal. Second, the motive or reason for violence is important. Acts that seem justified or morally correct can increase viewer aggression, whereas unjustified violence can actually decrease the risk of learning aggression. Third, the presence of weapons, especially guns and knives, can enhance aggression because such devices often trigger violent thoughts and memories in the viewer. Fourth, violence that is extensive or graphic can increase the risk of viewers learning aggression from a program. Exposure to extensive graphic violence also produces desensitization and can increase fear among viewers, the other two harmful effects mentioned above. Fifth, portrayals of violence that seem realistic are more likely to increase viewer aggression than are unrealistic scenes. Realistic violence also can increase audience fear. However, this does not mean that cartoon or fantasy violence on television is harmless, as discussed below. Sixth, violence that is explicitly rewarded or that simply goes unpunished increases the risk of learning aggression, whereas violence that is condemned decreases that risk. In addition, violence that goes unpunished can elevate fear, particularly when it appears to be unjust or random. Seventh, the negative consequences of violence for the victim are an important contextual cue. Showing the pain and suffering that result from violence can discourage the learning of aggression among viewers. Finally, violence that is cast in a humorous light can contribute to viewer aggression by making it seem rewarding. Humorous violence also may desensitize viewers to the seriousness of such behaviors.

Given the importance of context, NTVS created measures of all eight of these features so they could be evaluated anytime a violent portrayal was found on television.

Content Analysis

Acontent analysis is a scientific method for studying the features of different communication messages. Researchers can use a content analysis to analyze messages in newspapers, on the radio, on television, and even in people's diaries. Any recorded document can be studied by this method.

As an example, Roger Johnson (1999) was interested in looking at the types of stories featured on television news. He decided to study four different types of news programming: national network news (CBS), local news (a CBS affiliate), cable network news (CNN), and independent superstation news (WWOR, New York). The first step in a content analysis is to draw a representative sample of the documents to be studied. If Johnson had chosen a single night from each type of programming, his sample would have been limited to a certain set of stories or events occurring on that particular day. Instead, he randomly selected one newscast a week over a six-month period for each type of programming, resulting in a total sample of one hundred programs. In this way, Johnson's sample covers a range of time periods and possible news stories.

Next, the researcher must precisely define the measures and categories that will be analyzed within the documents. In this case, Johnson specified six categories of news stories: violent crime, tragedy, nonviolent conflict, social protest, war/military affairs, and all others. Two coders were trained to recognize these themes. Then, working independently, the coders watched the videotaped programs and placed each of the 1,798 separate news stories into one of the six categories. By having the two coders work alone, Johnson made sure that the judgments were as objective as possible. What did he find? Slightly more than half of all news stories (53%) featured violent crime, tragedy, or conflict, leading Johnson to conclude that news programming is dominated by "bad" news.

Content analyses differ in how good or scientifically valid they are. A valid content analysis should include a representative sample of documents, precise definitions of the measures, and independent coders who show a high level of agreement on their judgments. Johnson's study is characterized by all of these elements. As a contrast, compare it to a hypothetical study that looks only at network news during a particular week and that involves a single person judging the content. This may be how viewers themselves draw opinions about the nature of television news, but it is not very scientific or generalizable.

A content analysis can provide information about what is contained in mass-media messages, and it can even be used to compare the mass media with the real world to see how realistic or distorted the messages are. It cannot, however, be used to test the effect of media messages on the audience. Quite simply, it is not a study of how people react to the messages; rather, it is a study of the messages themselves.

Foundation 4

Both children and adults are influenced by the contextual features described above. Nevertheless, some unique concerns arise with regard to young children, particularly those under the age of seven years. Young children's cognitive abilities are still developing, so they often have interpretations of the television messages that are different from the interpretations of mature viewers. For example, studies by Peter Nikken and Allerd Peeters (1988) and by John Wright and his colleagues (1994) show that young children often have difficulty distinguishing reality from fantasy on television. In other words, what seems unrealistic to an older child or adult may be quite real to a preschooler. This helps to explain why researchers such as Lynette Friedrich and Aletha Stein (1973) found that young children will readily imitate violent cartoon characters.

Younger children also are more vulnerable to television violence in general. In a meta-analysis of 217 studies, Haejung Paik and George Com-stock (1994) found that compared to older viewers, preschoolers showed the strongest effects of television violence on aggressive behavior. Because of these age differences, NTVS identified younger children as a special audience when monitoring the content of television and reporting the findings.

Definition of Violence

Most people would agree that when a character shoots someone with a rifle, it is an example of violence. But what if they found out that the character tripped over a rock while hunting and the shooting was accidental? Violence can be defined in numerous ways, with each decision about the definition having important implications for the findings of a given study. For example, should verbal assaults be considered as violence? Should physical aggression directed at animals be included? What about humorous depictions of violence, such as slapstick?

Obviously, there are no "true" or correct answers to these questions. Yet researchers must clearly specify what gets "counted" as violence and what does not before embarking on a content analysis. According to NTVS, three key features must be present for a portrayal to qualify as violence: (1) the perpetrator as well as the target of violence must be animate or living beings, (2) there must be a clear intent to harm, and (3) the harm must be physical in nature as opposed to psychological or emotional. The following is an even more precise definition:

[Violence is] any overt depiction of a credible threat of physical force or the actual use of such force intended to physically harm an animate being or group of beings. Violence also includes certain depictions of physically harmful consequences against an animate being or group that occurs as a result of unseen violent means [National Television Violence Study, 1997, p. 41].

Thus, three forms of violence were included in the study: credible threats, behavioral acts, and harmful consequences of unseen violence.

Sample

One of the hallmarks of the NTVS project is its sample, which is substantially larger than that of any other single content analysis of television violence. Previous studies of television violence typically analyzed anywhere from 80 to 120 hours of programming. In contrast, NTVS videotaped about 2,700 hours of material each year during the three-year period, resulting in a total sample of more than 8,000 hours. Furthermore, most of the earlier studies examined programming on the three major broadcast networks only (ABC, NBC, and CBS). Although this may have reflected American viewing patterns in the past, more than two-thirds of American homes had cable television by 1995. To more fully capture the universe of American television in the 1990s, NTVS looked at programming across twenty-three different broadcast and cable channels.

Another strength of the sample is that instead of choosing intact days or weeks of programming, NTVS selected each individual program randomly from a population of all programs appearing across a nine-month timeframe. This approach ensured that an unusual news event, such as a schoolyard shooting or the breakout of war, would have much less effect on the overall representativeness of the sample. If an entire day or week had been chosen, such atypical news events could have dominated and therefore contaminated the sample.

Each year from October to June, the researchers randomly selected programs on twenty-three television channels to create a composite week of content for each channel. Programs were sampled between the hours of 6:00 A. M AND

11:00 p.m., across all seven days of the week. The twenty-three channels were those most frequently viewed by the American public, with the exception of CNN and ESPN, which were excluded because sports and breaking news were not evaluated in the study. The channels were grouped into five categories: broadcast network (ABC, NBC, CBS, and Fox), independent broadcast (Los Angeles-based KCAL, KCOP, and KTTV), public broadcasting (PBS), basic cable (A&E, AMC, Cartoon Network, Disney, Family Channel, Lifetime, Nickelodeon, TNT, USA, VH-1, and MTV), and premium cable (Cinemax, HBO, and Showtime).

In total, the 2,700 hours of television sampled each year resulted in more than 9,500 programs observed across the three-year period. As stipulated in the research contract with the study's funder (NCTA), all types of programs were analyzed for violence except for religious programs, game shows, infomercials, instructional shows, and breaking news. These five categories represented less than 15 percent of the programming each year.

Measuring Violence across Incidents, Scenes, and Programs

NTVS measured violence across three distinct levels or units of analysis. The "violent incident," the smallest unit, was defined as a violent interaction between a perpetrator and a victim. The "violent scene" was defined as a series of related violent incidents that occur without a significant break in the flow violence, such as a bar fight. The largest level of analysis was the entire "violent program." Some contextual features of violence were judged at the level of each violent incident, such as whether the perpetrator was attractive. Other features were evaluated by taking the entire violent scene into account, such as how graphic was the violence. Still other features required a consideration of the entire program as a unit, such as whether there was an overall theme of antiviolence in the show. By analyzing violence at all three levels, the researchers were able to provide comprehensive information about the meaning of violence on television.

Coding and Training

An elaborate codebook was developed to provide detailed and precise definitions of terms such as "credible threat," "rewarded violence," and "long-term consequences." The codebook also provided extensive rules of judgment for coding. Each year, more than fifty undergraduate students at the University of California, Santa Barbara, were trained to become highly skilled at applying the definitions and rules laid out in the codebook. These coders received approximately forty hours of classroom instruction and twenty hours of laboratory practice to help them learn the complex coding scheme. Once trained, the coders worked individually in quiet laboratories as they assessed programs for violent content.

It took approximately twenty weeks each year to complete all the coding. Every week during this time, half of the coders independently evaluated the same program to assess consistency across individuals. Agreement or reliability was at least 80 percent on nearly all the measures coded each year.

Major Findings

The NTVS researchers found a remarkable degree of consistency in how violence was portrayed across the three years of the study. They also found that the way that most television violence is portrayed does in fact pose risks to viewers. What are some of the portrayals that can be harmful?

First, much of television violence is "glamorized," or cast in a positive light. Across three years of the study, nearly 40 percent of the violent incidents on television were initiated by "good" characters who can serve as attractive role models (see Figure 1). The risk here is that viewers of all ages, but especially children, are likely to emulate characters who are perceived as attractive.

Another aspect of glamorization is that physical aggression is often condoned on television. For example, more than one-third of violent programs featured bad characters who are never punished anywhere in the plot. Violence that goes unpunished poses risk because it can encourage the learning of aggression among viewers. In addition, fully 71 percent of violent scenes contained no remorse, criticism, or penalty for violence at the time that it occurs. This is especially problematic for viewers under the age of seven years who focus mostly on immediate repercussions and lack the cognitive ability to consider punishments for aggression that may occur later in a program.

Second, most television violence is "sanitized," or shown with minimized negative consequences. In fact, roughly half of the violent incidents on television showed no physical harm or pain to the victim (see Figure 2). Not only are short-term outcomes often missing, so are longterm consequences. Over the three-year period, only 15 percent of the violent programs showed any prolonged negative effects of violence on the family, friends, or community of the victim. As discussed above, the portrayal of negative outcomes such as pain and suffering can decrease the chances that viewers will learn aggression from media violence.

Third, much of the serious violence on television is "trivialized," or shown as less serious than it is. Across the three-year study, more than half of the violent incidents featured physical aggression that would be deadly or incapacitating if it were to occur in real life (see Figure 3). Yet much of this serious violence is undercut by humor. In fact, about 40 percent of the violent scenes on television are shown in a humorous context. Exposure to serious violence that is made to seem trivial can contribute to both desensitization and imitation among viewers.

Fourth, very few programs emphasize an antiviolence theme. A program can include violence in a way that is actually educational rather than promotional of violence. For example, violence can be shown to have strong negative consequences or alternatives to violence can be promoted. Less than 5 percent of violent programs featured an antiviolence message across the three years of the study.

Thus, what is the prevailing message? Violence on American television is typically presented as an acceptable and often beneficial way to solve problems, and it rarely results in any serious or lasting damage.

The NTVS researchers also looked at the sheer prevalence of violence, independent of how it is portrayed. Over the three years of the study, a steady 60 percent of programs across the television landscape contained some violence. However, the prevalence of violence varied quite a bit across the different channel types, with premium cable the most likely to feature violent programs (more than 80% of the programs on these channels contained violence) and public broadcasting the least likely (fewer than 20% of the programs on this channel contained violence) (see Figure 4).

The researchers also looked at the concentration or amount of violence within a program. They found that the typical violent program featured at least six violent incidents per hour. Each of these incidents involved a different perpetrator and victim, and many of them included multiple acts of physical aggression. In fact, more than 60 percent of the violent incidents contained repeated behavioral acts of aggression on a victim. In other words, it is rare for a television perpetrator to hit, shoot, or stab someone only once.

The last major finding concerns what the NTVS researchers called "high-risk" portrayals. Depictions were labeled "high risk" when several plot elements that encourage the learning of aggression are all featured in one scene. A high-risk portrayal features an attractive perpetrator, engaging in justified violence that goes unpunished, that results in minimal consequences to the victim, and that seems realistic to the viewer.

The researchers found that for children under the age of seven years, high-risk portrayals of violence that teach aggression are found most often in cartoons, the very programs that are targeted to this age group. Of course, cartoons pose little risk for older, more mature viewers who routinely discount animated content as unrealistic. But preschoolers have difficulty distinguishing reality from fantasy on television, and therefore are susceptible to imitating fantasy violence that occurs on television.

Taking into account the typical viewing habits of children in the United States, the researchers concluded that the average American preschooler who watches mostly cartoons is exposed to more than five hundred high-risk portrayals of violence each year. These portrayals contain a potent set of plot features that strongly encourage the learning of aggression in some of the most vulnerable members of the television audience: young children.

Implications

The NTVS received a great deal of public attention. The findings each year made headline news in major newspapers such as The New York Times, The Washington Post, USA Today, and the Los Angeles Times. In addition, the study was featured on CNN, on the ABC, CBS, and NBC news programs, and on the "Jim Lehrer Online NewsHour."

The study also was recognized by prominent policymakers. Senator Paul Simon (D-IL) described it as "a solid report that should be a strong signal to the television industry and to the public that glorified violence is harmful and that we still need improvement" (Simon, 1996). Representative Joseph P. Kennedy (D-MA) stated that "The National Television Violence Study shows the problem is not going away. The television industry must work together and clean up media violence for the sake of public health—our children's health" (Kennedy, 1997). Even the president of the United States acknowledged the study. The researchers briefed the White House regarding the findings, and President Bill Clinton issued a statement noting that the study "shows that there is still too much violence on the television, and that we must all continue the important work we have begun" (Clinton, 1997).

However, the effect of the study was not limited to political rhetoric alone. In 1996, Representative Edward Markey (D-MA) argued that NTVS was instrumental in helping to pass V-chip legislation, mandating that a blocking device be installed in new television sets. Markey (1996, pp. 5-6) described the study's report as "devastating" for the television industry, noting the following:

It shredded the industry argument that the V-chip was unnecessary because voluntary action was working. Instead, the authors found that in the programs studied, perpetrators were going unpunished in 73% of all violent scenes, and only 4% of violent programs emphasized anti-violence themes.… This may have been the greatest paradox of all—that a study initiated in 1994 to block passage of the V-chip should become one of the most powerful arguments to go forward with the V-chip in 1996.

Beyond policymaking, the study has had an effect on science as well. The American Psychological Association (1998) called it "the gold standard against which other research on television violence will be measured." George Comstock and Erica Scharrer (1999) described NTVS as "the most powerful content analysis in the history of mass communication research." And Steven Chaffee (1999), a distinguished scholar of media effects, described the content analysis as "an exemplary model for researchers for years to come." Already, researchers in Great Britain and Portugal have used the NTVS methods and coding scheme to analyze television violence in their own countries.

Conclusion

The ultimate measure of the effect of any study is whether it has long-term scientific and practical significance. If NTVS encourages future researchers of television violence to focus not so much on the number of shootings, stabbings, and fistfights, but instead on the meaning of this aggression within the context of the plot, then it will have lasting scientific value. If the study helps parents to navigate the television landscape and make informed choices about their children's viewing habits, then it will have practical significance as well.

Bibliography

American Medical Association. (1996). Physician Guide to Media Violence. Chicago: American Medical Association.

American Medical Association (1998). "AMA Says New Study 'Definitive Confirmation' That TV Violence Is Harmful." Press release from the American Medical Association, April 16.

American Psychological Association. (1993). Violence and Youth: Psychology's Response. Washington, DC: American Psychological Association.

Chaffee, Steven. (1999). "National Television Violence Study 3." Journal of Communication 49(3):194-196.

Clinton, William J. (1997). "Statement by President Clinton on the Release of the second Annual National Television Violence Study." Washington, DC: The White House Office of the Press Secretary, March 28.

Comstock, George, and Scharrer, Erica. (1999). Television: What's On, Who's Watching, and What it Means. New York: Academic Press.

Friedrich, Lynette K., and Stein, Aletha H. (1973). "Aggressive and Prosocial Television Programs and the Natural Behavior of Preschool Children." Monographs of the Society for Research in Child Development 38(4, Serial No. 151).

Getlin, Josh. (1999). "Grim Show Is Only the Latest for America's Violent Culture." Los Angeles Times, April 22, p. A17.

Hoffner, Cynthia. (1998). "Framing of the Television Violence Issue in Newspaper Coverage." In Television Violence and Public Policy, ed. James T. Hamilton. Ann Arbor: University of Michigan Press.

Huesmann, L. Rowell. (1986). "Psychological Processes Promoting the Relation between Exposure to Media Violence and Aggressive Behavior by the Viewer." Journal of Social Issues 42(3):125-140.

Johnson, Roger N. (1996). "Bad News Revisited: The Portrayal of Violence, Conflict, and Suffering on Television News." Peace and Conflict: Journal of Peace Psychology 2(3):201-216.

Kennedy, Joseph P., II. (1997). "The TV Industry Must Work Together and Clean Up Violence." Press release from the Office of Representative Joseph P. Kennedy II (D-MA), March 26.

Markey, Edward J. (1996). "It Takes a V-Chip." <http://www.house.gov/markey/chip.htm>.

National Institute of Mental Health. (1982). Television and Behavior: Ten Years of Scientific Progress and Implications for the Eighties (Vol. 1). Summary Report. Washington, DC: U.S. Government Printing Office.

National Television Violence Study, 1. (1997). Thousand Oaks, CA: Sage Publications.

National Television Violence Study, 2. (1998a). Thousand Oaks, CA: Sage Publications. National Television Violence Study, 3. (1998b). Thousand Oaks, CA: Sage Publications.

Nikken, Peter, and Peeters, Allerd L. (1988). "Children's Perceptions of Television Reality." Journal of Broadcasting and Electronic Media 32:441-452.

Paik, Haejung, and Comstock, George. (1994). "The Effects of Television Violence on Antisocial Behavior: A Meta-Analysis." Communication Research 21:516-546.

Pew Research Center. (1999). "Teens and Traffic Top Community Concerns." <http://www.people-press.org/may99rpt2.htm>.

Simon, Paul. (1996). "Reaction of Senator Paul Simon to the Cable-Funded Audit of TV Violence." Press release from the Office of Senator Paul Simon (D IL), February 6.

Wright, John C.; Huston, Aletha C.; Reitz, Alice Leary; and Piemyat, Suwatchara. (1994). "Young Children's Perceptions of Television Reality: Determinants and Developmental Differences." Developmental Psychology 30:229-239.

Barbara J. Wilson

Encyclopedia of Communication and Information