It is guns, it is poverty, it is overcrowding, and it is the uniquely American problem of a culture that is infatuated with violence. We love it, we glamorize it, we teach it to our children.l

The above testimony by Dr. Deborah Prothrow-Stith on gangs and youth violence presented before the U.S. Senate contains two important points concerning the mass media and youth violence. First, it does not mention the media as a factor in violence, lending support to the view that the media are not crucial agents in youth violence. Second, it does cite an American culture that is infatuated with violence, and the glamorization and teaching of violence to our children, as problems. Culture, glamorization, and instruction, however, are areas where the media have been shown to play important social roles. The above statement simultaneously provides support for the position that the media are indeed important players in the production of youth violence and yet paradoxically also supports the position that they are not contributors. The relative validity of these two dichotomous positions, the media as unimportant and the media as central in fostering youth violence, has dominated the public discussion, resulting in much confusion about this issue and public posturing by various groups and individuals. The actual relationship of the media to youth violence lies somewhere between these two extremes.

Research interest in the relationship of the mass media to social violence has been elevated for most of this century. Over the twentieth century, the issue of the media as a source of violence has moved into and out of the public consciousness in predictable ten- to twenty-year cycles. If a consensus has emerged from the research and public interest, it is that the sources of violence are complex and tied to our most basic nature as well as the social world we have created and that the media's particular relationship to social violence is extremely complicated. (See the discussion in this author's Media, Crime, and Criminal Justice [1992] and in Crime and Human Nature [1985] by J. Wilson and R. Herrnstein.)

Therefore, when discussing the nature of the relationship between the media and violence, it is important not to be myopic. Social violence is embedded in historical, social forces and phenomena, while the media are components of a larger information system that creates and distributes knowledge about the world. The media and social violence must both be approached as parts of phenomena that have numerous interconnections and paths of influence between them. Too narrow a perspective on youth violence or the media's role in its generation oversimplifies both the problem and the solutions we pursue. Nowhere is this more apparent than in the current concern about media, youth, and violence.

STATISTICS ON YOUTH VIOLENCE

The source of this concern is revealed by a brief review of the statistics of youth violence.2 Youth violence, and particularly violent crime committed by youth, has recently increased dramatically. Today about 5 out of every 20 robbery arrests and 3 of every 20 murder, rape, and aggravated assault arrests are of juveniles. In raw numbers, this translates into 3,000 murder, 6,000 forcible rape, 41,000 robbery, and 65,000 aggravated assault arrests of youths annually.

The surge in youth criminal violence is concentrated within the past five years. During the first part of the 1980s, there was a general decline in youth arrests for both violent and property crimes. In the latter half of the 1980s, however, youth arrests increased at a pace greater than that of adults for violent crimes. Youth arrests increased substantially between 1981 and 1990 for nonaggravated assault (72 percent), murder and nonnegligent manslaughter (60 percent), aggravated assault (57 percent), weapons violations (41 percent), and forcible rape (28 percent). Looking over a generational time span from 1965 to 1989, the arrest rate for violent crimes by youths grew between the mid-1960s and the mid-1970s but then leveled off and remained relatively constant until the late 1980s. At that time, the rate again began to increase, reaching its highest recorded level in the most recent years.

Thus, while the proportion of youth in the general population has declined as the baby-boom generation has aged, the rate of violence from our youth has increased significantly. We have fewer youth proportionately, but they are more violent and account for increased proportions of our violent crime. Attempts to comprehend and explain this change have led invariably to the mass media as prime suspects, but deciphering the media's role has not been a simple or straightforward task.

This difficulty in deciphering the media's role is due to the fact that the relationship of media to violence is complex, and the media's influence can be both direct and indirect. Research on their relationship (reported, for example, in George Comstock's 1980 study Television in America) has revealed that media effects that appear when large groups are examined are not predictable at the individual subject level. The media are also related to social violence in ways not usually considered in the public debate, such as their effects on public policies and general social attitudes toward violence.

Adding to the complexity of the media's relationship, there are many other sources of violence that either interact with the media or work alone to produce violence. These sources range from individual biology to characteristics of our history and culture. The importance of nonmedia factors such as neighborhood and family conditions, individual psychological and genetic traits, and our social structure, race relations, and economic conditions for the generation of violence are commonly acknowledged and analyzed, as in Jeffrey Goldstein's 1986 study Aggression and Crimes of Violence. The role of the mass media is confounded with these other sources, and its significance is often either lost or exaggerated. One task of this essay is thus to dispel the two popular but polarizing notions that have dominated the public debate. The first is that the media are the primary cause of violence in society. The second is that the media have no, or a very limited, effect on social violence.

The former view of the media as the source of primary effects is often advanced along with draconian policy demands such as extensive government intervention or direct censorship of the media. The counterargument to this position is supported by a number of points. The most basic is that we were a violent nation before we had mass media, and there is no evidence that the removal of violent media would make us nonviolent.3 Some research into copy-cat crime additionally provides no evidence of a criminalization effect from the media as a cause.4 The media alone cannot turn a law-abiding individual into a criminal one nor a nonviolent youth into a violent one. In sum, individual and national violence cannot be blamed primarily on the media, and violence-reducing policies directed only at the media will have little effect.

The latter argument, that the media have limited to no effect on levels of social violence, is structured both in posture and approach to the tobacco industry's response to research linking smoking to lung cancer and it rings just as hollow. The argument's basic approach is to expound inherent weaknesses in the various methodologies of the media-violence research and to trumpet the lack of evidence of strong, direct effects, while ignoring the persistent pattern of positive findings. Proponents of the nil effect point out that laboratory experiments are biased toward finding an effect. To isolate the effect of a single factor, in this case the media, and observe a rare social behavior, namely violence, experiments must exaggerate the link between media and aggression and create a setting that will elicit violent behavior. They therefore argue that all laboratory research on the issue is irrelevant. They continue, however, to dismiss the nonlaboratory research because of a lack of strict variable controls and designs that leave open noncausal interpretations of the results. "No effects" proponents lastly argue that while society reinforces some behaviors shown in the media such as that found in commercials, it does not condone or reinforce violence and, therefore, a violence-enhancing effect should not be expected (a view discussed in "Smoking Out the Critics," a 1984 Society article [21:36-40] by A. Wurtzel and G. Lometti).

In reality, the research shows persistent behavioral effects from violent media under diverse situations for differing groups.5 Regarding the strong behavioral effects apparent in fashion and fad, effects that Madison Avenue touts, the argument of a behavioral effect only on sanctioned behavior but not on unsanctioned violence is specious. The media industry claim of having only positive behavioral effects is as valid as the tobacco industry claiming that their ads do not encourage new smokers but only persuade brand switching among established smokers. First, violence is sometimes socially sanctioned, particularly within the U.S. youth and hyper-masculine culture that is the target audience of the most prominently violent media: And although the media cannot criminalize someone not having criminal predispositions, media-generated, copy-cat crime is a significant criminal phenomenon with ample anecdotal and case evidence providing a form for criminality to take.6 The recurring mimicking of dangerous film stunts belies the argument of the media having only positive behavioral effects. It is apparent that while the media alone cannot make someone a criminal, it can change the criminal behavior of a predisposed offender.

CONFLICTING CAUSAL CLAIMS

The two arguments of primary cause and negligible cause compete for public support. These models not only posit differing causal relations between the media and violence but imply vastly different public policies regarding the media as well. The primary-cause model is that of a significant, direct linear relationship between violent media and violent behavior. In this model, violent media, independent of other factors, directly cause violent behavior. If valid, it indicates that strong intervention is necessary in the content, distribution, and creation of violent media.

The negligible-cause model concedes a statistical association between the media and violent behavior but poses the connection as due not to a causal relationship but to persons predisposed to violence simultaneously seeking out violent media and more often behaving violently. As the relationship is associative and not causal in this model, policies targeted at the media will have no effects on violent behavior and the media can be safely ignored.

Both models inaccurately describe the media-violence relationship. The actual relationship between the factors is felt to be bidirectional and cyclical. In addition to violently predisposed people seeking out violent media and violent media causing violent behavior, violent media play a role in the generation of violently predisposed people through their effects on attitudes. And as the made-for-TV movie industry reflects, violent behavior sometimes results in the creation of more violent media. Finally, by providing live models of violence and creating community and home environments that are more inured to and tolerant of violence, violent behavior helps to create more violently predisposed youth in society. Therefore, while the direct effect of media on violence may not be initially large, its influence cycles through the model and accumulates.

An area of research that provides an example of the bidirectional model is the relationship of pornography to sexual violence; a recent (1993) overview of such research can be found in Pornography, by D. Linz and N. Malamuth. On one hand, the research establishes that depictions of sexual violence, specifically those that link sex with physical violence toward women, foster antisocial attitudes toward women and lenient perceptions of the crime of rape. Aberrant perceptions, such as increased belief in the "rape myth" (that women unconsciously want to be raped or somehow enjoy being raped), have been reported. Virtually none of the research, however, reveals strong direct effects from pornography, and even sexually violent media do not appear to negatively affect all male viewers. Many cultural and individual factors appear to mediate the effects and to foster the predisposition to sexually violent media and sexual violence. Researchers in this area have concluded that the media are one of many social forces that affect the development of intervening variables, such as thought patterns, sexual arousal patterns, motivations, and personality characteristics that are associated with tolerance for sexual violence and perhaps an increase in sexually violent behavior in society.7 As in other areas of media-violence research, sexually violent media emerge as neither a primary engine nor an innocuous social factor.

THE KEY TO MEDIA EFFECTS

The key to media effects occurring in any particular instance, then, are the intermediate, interactive factors. In terms of the media, there are numerous interactive factors that have been identified as conducive to generating aggressive effects. Among the many delineated in the research, a sample includes: reward or lack of punishment for the perpetrator, portrayal of violence as justified, portrayal of the consequences of violence in a way that does not stir distaste, portrayal of violence without critical commentary, the presence of live peer models of violence, and the presence of sanctioning adults (all discussed in Comstock's Television in America). Only unambiguous linking of violent behavior with undesirable consequences or motives by the media appears capable of inhibiting subsequent aggression in groups of viewers.

A list of nonmedia factors deemed significant in the development of crime and the number of violently predisposed individuals can be culled from Crime and Human Nature by J. Wilson and R. Herrnstein. The authors list constitutional, developmental, and social-context factors including gender, age, intelligence, personality, psychopathology, broken and abusive families, schools, community, labor markets, alcohol and heroin, and finally history and culture. As can be seen, most aspects of modern life are implicated, and only tangential factors like diet and climate (which other researchers would have included) are left out. With such a large number of factors coming into play, the levels of interactions and complexity of relationships are obviously enormous.

The research on violence suggests that certain factors are basic to violent crime, as detailed by Wilson and Herrnstein. None of these factors dominates, but none are without significant effects.

Accordingly, the research (contained in this author's 1992 study Media, Crime and Criminal Justice) clearly signifies the media as only some of many factors in the generation of youth violence and that media depictions of violence do not affect all persons in the same way. The media contribute to violence in combination with other social and psychological factors. Whether or not a particular media depiction will cause a particular viewer to act more aggressively is not a straightforward issue. The emergence of an effect depends on the interaction between each individual viewer, the content of the portrayal, and the setting in which exposure to the media occurs. This gives the media significant aggregate effects but makes these effects difficult to predict for individuals. There is no doubt, however, that violent children, including those who come to have significant criminal records, spend more time exposed to violent media than do less violent children. The issue is not the existence of a media effect but the magnitude or importance of the effect.

Media violence correlates as strongly with and is as causally related to the magnitude of violent behavior as any other social behavioral variable that has been studied. This reflects both the media's impact and our lack of knowledge about the etiology of violence. Because of the many individual and social factors that come into play in producing any social behavior, one should not expect to find more than a modest direct relationship between the media and violence. Following their review of the research, Thomas Cook and his colleagues conclude:

No effects emerge that are so large as to hit one between the eyes, but early measure of viewing violence adds to the predictability of later aggression over and above the predictability afforded by earlier measures of aggression. These lagged effects are consistently positive, but not large, and they are rarely statistically significant, although no reliable lagged negative effects have been reported. . . . But is the association causal? If we were forced to render a judgment, probably yes. . . . There is strong evidence of causation in the wrong setting (the lab) with the right population (normal children) and in the right setting (outside the lab) with the wrong population (abnormal adults).8

MEDIA AMONG MANY FACTORS

In summation, despite the fact that the media are among many factors, they should not be ignored, regardless of the level of their direct impact. Because social violence is a pressing problem, even those factors that only modestly contribute to it are important. Small effects of the media accumulate and appear to have significant long-term social effects.9 The research strongly indicates that we are a more violent society because of our mass media. Exactly how and to what extent the media cause long-term changes in violent behavior remains unknown, but the fact that it plays an important, but not independent, role is generally conceded.

What public policies are suggested by the knowledge we now possess about media and violence? Not all of the factors discussed above are good candidates for public intervention strategies, but there are three sources of youth violence that government policy can influence. In order of importance, they are: extreme differences in economic conditions and the concentration of wealth in America; the American gun culture; and, exacerbating the problems created by the first two, the media's violence-enhancing messages. Family, neighborhood, and personality factors may be more important for generating violence in absolute magnitude, but they are not easily influenced by public actions.

The magnitude of economic disparity and the concentration of wealth in the United States is greater than in comparable (and, not surprisingly, less violent) societies. Our richest citizens not only earn vastly more than our poorest, but, more important, the wealth in the country is increasingly concentrated in fewer and fewer hands. The trend during this century, which accelerated during the 1980s, is for an ever-shrinking percentage of the richest Americans to control greater proportions of the country's wealth, while the poorest have access to increasingly smaller proportions. The burden of this economic disenfranchisement, both psychologically and fiscally, falls heavily on the young, and especially on the young who are urban poor minorities, as is shown In Elliott Currie's 1985 study Confronting Crime. In a consumerism saturated society like the United States, hopelessness, bitterness, and disregard for moral values and law are heightened by this growing economic disparity.

And as the economic polarization and violent crime have grown, we also became nationally fixated on heightening and extending our punishment capacities in an attempt to suppress violent behavior, evidenced by Diana Gordon's 1991 study The Justice Juggernaut. Since 1975, we have increased the rate of juvenile incarceration steadily. Today we hold in custody approximately one hundred thousand juveniles every year. Despite our strengthened capacity to punish, however, youth violence has not abated.

This result should have been expected because two social mechanisms are needed to reduce violence--punishing violent criminal behavior and rewarding law-abiding, nonviolent behavior. Societies that are more successful in balancing the two mechanisms are less violent, as shown in Crime and Control in Comparative Perspective, by H. Heiland and L. Shelley (1992). While punishment of violent behavior is certainly necessary and justified, its emphasis, coupled with the concentration of wealth in America, has resulted in the degrading of the equally important social capacity to reward law-abiding behavior. By emphasizing one, we have lamed and discredited the other. Nonmaterial rewards like social status, an esteemed reputation, and a clear conscience have been losing their legitimacy with the young, while material rewards for law-abiding life-styles such as careers, comfortable incomes, and affordable goods are less generally available to our poorest and, not surprisingly, most crime-prone and violent citizens.

We have chosen to emphasize the mechanism, punishment, that is actually the weaker of the two in actually influencing behavior. As operant conditioning theory would predict, punishment, if severe enough, can suppress one type of violent crime. But the suppression of one behavior gives no push toward a desirable replacement activity, and a substitute violent crime will likely emerge. So "smash and grab" robberies give way to "bump and rob" holdups. Shaping behavior requires a credible reward system. In social terms, youth must see law-abiding behavior as credible and potentially rewarding as well as seeing violent behavior as potentially resulting in punishment.

The second area that government policy can immediately address is the gun culture in America. Our culture of violence, referred to in the opening quote, is made immeasurably more deadly by the enfolded gun culture. The availability of guns as cheap killing mechanisms is simply a national insanity. The mass production of these killing "toys" and the easy access to them must be addressed. The most recent statistics show that one out of every ten high school students report that they carry a handgun. Gun buy-back programs should be supported, and production and availability must be reduced if a positive net effect is to be expected. Irrespective of the difficulty of controlling the sources of individual violent behavior, the implements of fatal violence should not be ignored.

The third area of policy concern, the mass media, exacerbates the gun culture by portraying guns as glamorous, effective, omnipotent devices. The mass media also heighten the negative effects of economic disparities through their consumer messages in advertising and entertainment. Although both of these effects that add to the problem of youth violence are sometimes discussed, the debate about the media remains tightly focused on measuring and reviewing violent media content. Within this focus, the emphasis has been on counting violent acts rather than on exploring the context of its portrayal. Deciphering the media's moral and value messages about violence has been mostly ignored.

EFFECTS ON CRIMINAL JUSTICE

A closer examination of the context of violence in the media would tell us that we should not try to purge the media of violence, for violent media can be good when programs teach that violence is bad. Our goal should be to reduce graphic, gratuitous, and glorified violence; to portray it not as a problem solver but as a reluctant, distasteful, last resort with tragic, unanticipated consequences. Violence shown consistently as a generator of pain and suffering, not as a personal or social panacea, would be positive media violence. Too often, violence in the media is shown as an effective solution, and, too often, it is simply met by increased counterviolence. But, despite the recurring interest and current debate about media violence, there is little direction for the media industry regarding the context of violence and its effects. A goal should be to provide better information to the industry that details the various contexts and messages of violence and their effects.

Perhaps the most significant social effect of media violence is, however, not the direct generation of social violence but its impact on our criminal justice policy. The fear and loathing we feel toward criminals--youthful, violent, or not--is tied to our media-generated image of criminality. The media portray criminals as typically animalistic, vicious predators. This media image translates into a more violent society by influencing the way we react to all crime in America. We imprison at a much greater rate and make reentry into law-abiding society, even for our nonviolent offenders, more difficult than other advanced (and, not coincidentally, less violent) nations. The predator-criminal image results in policy based on the worst-case criminal and a constant ratcheting up of punishments for all offenders. In its cumulative effect, the media both provide violent models for our youth to emulate and justify a myopic, harshly punitive public reaction to all offenders.

Currently, the debate concerning both the media and youth violence has evolved into "circles of blame" in which one group ascribes blame for the problem to someone else in the circle. Thus, in the media circle, the public blames the networks and studios, which blame the producers and writers, who blame the advertisers, who blame the public. In the violence circle, the government blames the youth, who blame the community, which blames the schools, which blame the parents, who blame the government. A more sensible, productive process would be a shift; to a "ring of responsibility," with the groups addressing their individual contributions to the problem and arriving at cooperative policies. We can't selectively reduce one aspect of violence in a violent society and expect real results. Youth violence will not be seriously reduced without violence in other aspects of our culture being addressed. In the same vein, modifying media violence alone will not have much effect but to ignore it will make efforts on other fronts less successful. Ironically, despite the fact that the media have limited independent effects on youth violence, we need to expand the focus on them. This should incorporate other social institutions, such as the media industry itself, and the social norms and values reflected in the media. We could then derive more general models of media effects and social violence.

Violence is a cultural product. The media are reflections of the culture and engines in the production process. Although they are not the only or even the most powerful causes, they are tied into the other violence-generating engines, and youth pay particular attention to them. The aggregate result of all of these forces in the United States is a national character that is individualistic, materialistic, and violence prone. If we wish to change our national character regarding violence, we cannot take on only some aspects of its genesis. We must address everything we can, such as economic inequities, the gun culture, and the glamorization of violence. And, by a slow, painful, generational process of moral leadership and example, we must work to modify the individual, family, and neighborhood factors that violently predispose youth.

In conclusion, our youth will be violent as long as our culture is violent. The local social conditions in which they are raised and the larger cultural and economic environments that they will enter generate great numbers of violently predisposed individuals. As we have experienced, violently predisposed youth, particularly among our poor, will fully develop their potential and come to prey upon us. Faced with frightful predators, we subsequently and justly punish them, but the use of punishment alone will not solve the problem. The role that the media play in the above scenario versus their potential role in deglorifying violence and showing our youth that armed aggression is not an American cultural right, will determine the media's ultimate relationship to youthful violence in society.

ADDITIONAL READING

George Comstock, Television in America, Sage, Newbury Park, Calif., 1980.

Elliott Currie, Confronting Crime, Pantheon, New York, 1985.

Jeffrey Goldstein, Aggression and Crimes of Violence, Oxford University Press, New York, 1986.

Diana Gordon, The Justice Juggernaut, Rutgers University Press, New Brunswick, N.J., 1991.

Joshua Meyrowitz, No Sense of Place, Oxford University Press, New York, 1985.

Ray Surette, Media, Crime, and Criminal Justice, Brooks/Cole, Pacific Grove, Calif., 1992.

James Q. Wilson and Richard Herrnstein, Crime and Human Nature, Simon & Schuster, New York, 1985.

1. Dr. Deborah Prothrow-Stith testifying before the Senate Subcommittee on Juvenile Justice, November 26, 1991.

2. Sources of the statistics cited in this essay are drawn from "Arrests of Youth 1990," January 1992, Office of Juvenile Justice and Delinquency Prevention Update on Statistics; and Sourcebook of Criminal Justice Statistics--1992, Bureau of Justice Statistics, U.S. Department of Justice, 1993.

3. Hugh Davis Graham and Ted Robert Gurr, eds., Violence in America (Beverly Hills, CA: Sage, 1979).

4. See S. Milgram and R. Shotland, Television and Antisocial Behavior Field Experiments (New York: Academic Press, 1973) and A. Schmid and J. de Graaf, Violence as Communication (Newbury Park, CA: Sage, 1982).

5. T. Cook, D. Kendzierski, and S. Thomas, "The Implicit Assumptions of Television Research," Public Opinion Quarterly 47:161-201.

6. For a listing of examples see S. Pease and C. Love, "The Copy-Cat Crime Phenomenon," in Justice and the Media by R. Surette (Springfield, IL: Charles C. Thomas, 1984), 199-211; and A. Schmid and J. de Graaf, Violence as Communication (Newbury Park, CA: Sage, 1982).

7. N. Malamuth and J. Briere (1986), "Sexual Violence in the Media: Indirect Effects on Aggression against Women," Journal of Social Issues 42, 89.

8. T. Cook, D. Kendzierski and S. Thomas (1993),"The Implicit Assumptions of Television Research," Public Opinion Quarterly 47:191-92.

9. R. Rosenthal (1986), "Media Violence, Anti-Social Behavior, and the Social Consequences of Small Effects," Journal of Social Issues 42:141-54.