RHODES: It always seemed to me intuitively wrong that media exposure influenced violent development. Violence varies greatly by ethnic group in the United States, for example, even though we're all exposed to comparable levels of mock violence in US media. Mock violence is much more extensive in Japanese media than in US but Japan has some of the lowest rates of criminal violence in the world. But until I encountered the research of American criminologist Dr. Lonnie Athens, who shows causally (not merely correlationally) that serious violent behavior is always the result of having been violently socialized, I had no evidence to back up my intuition. Athens's work supplied that evidence, as I tried to demonstrate in my 1999 book Why They Kill.

I was especially offended by the national political reaction to the Columbine school shootings. To blame that ugly event on the media, make public pronouncements of concern, endorse dubious research and leave the field triumphant seemed to me little short of obscene when the basis of such violence, as Athens's work demonstrates, is society's failure to prevent the brutalization of children and failure to intervene with support and help when those brutalized children begin moving toward violence as an answer to their lack of protection.

So I decided to investigate the media violence research claims armed with Athens's model. It was no surprise to discover that they have essentially no evidential support. I was surprised, however, to find poorly conceived, scientifically inadequate, biased and sloppy if not actually fraudulent research.

ABFFE: Why has condemning media violence become so popular among politicians?

RHODES: I can't speak to individual motive, but it's clear at least that media violence is a popular issue; that many people would like to believe that it predisposes children to misbehavior (following in an old tradition that once blamed the novel for subverting young women's virtue, theater for corrupting the young, comic books for teaching criminal attitudes and skills--the old tradition, in short, that one generation's trash is the next generation's art form) and are happy to have some marginal social science seem to prove them right; that politicians can embrace the issue and win votes knowing that the First Amendment prevents them from having to do anything about it. I have even heard politicians say--off the record--that it's an ideal smokescreen to protect them from having to confront the much more controversial question of gun control.

ABFFE: Politicians claim that the scientific evidence that media violence is harmful is as strong as the evidence that smoking causes cancer. Is that true?

RHODES: The supposed analogy between smoking/cancer and media violence/violence was invented, so far as I can tell, by Rowell Huesmann, a professor of psychology at the University of Michigan and the most prominent current exponent of the media violence theory. I discuss his work at length in my report "The Media Violence Myth." Huesmann bases his claim for a similarity on the numerical similarity of the correlations. But in fact the similarity stops there, at the trivial level of two similar numbers.

Smoking is an identifiable and quantifiable behavior. Lung cancer is a discrete biological entity which any trained observer can recognize and which is the outcome of a series of specific biological processes. Media violence is a social construct that depends entirely on what one group or individual or another defines it to be. Huesmann considers Three Stooges comedies to be violent, whereas most of us consider them to be funny. He considers Roadrunner cartoons to be extremely violent, whereas most of us, including four-year-olds, recognize that they're comedic fantasies with no direct connection to the real world. One of the most glaring omissions in so-called media violence studies is athletic events. Football, hockey and other sports fill hours of air and cable time, yet none of the studies clock children's exposure to their real, not mock, violence. So what is media violence? Not, certainly, a clearly defineable concept.

Nor is "violence" the measure that Huesmann and others have generally used in their research; rather, they study what they call "aggression," which they define differently from one study to the next if they define it at all. Sometimes they seem to be talking about verbal aggression, sometimes physical, and the physical may vary from posturing to pushing to hitting. In the real world, of course, aggression in many contexts is considered to be a virtue, not a vice--ask political candidates and corporate CEOs whether aggression is a virtue or a social problem.

Finally, the questionable correlation Huesmann claims to have found between media violence and violence is based on a sample of a few hundred people. The correlation between smoking and cancer is far more robust, with tens and hundreds of thousands of cases in the sample.

Smoking/cancer and media violence/violence are not comparable correlations.

ABFFE: Why are some social scientists defending the media violence hypothesis if the evidence is so weak?

RHODES: Social scientists are not immune from wishful thinking, ambition, inadequate logical skills and yearnings to ride popular bandwagons in the great morality parade. It's very difficult to study group behavior in human beings. Correlational studies are about the best that can be done, and correlations, by definition, do not reveal causes. They only reveal interesting patterns that may (or may not) lead researchers to causes. That means the social sciences depend much more heavily on interpretation than the hard sciences, opening a broad playing field for mischief, wish fulfillment and deliberate distortion. I found much that I believe to be unethical and even fraudulent in the media violence research I investigated.

ABFFE: What are the most useful ways of reducing youth violence?

RHODES: Athens's identification of violent socialization as the cause of violent criminality--using a research methodology that does locate causes and effects--makes it possible to prevent or interdict violent development. Socialization toward violent criminality begins with brutalization, usually in childhood (meaning someone violently dominates a child, the child sees loved ones violently dominated, and the child is coached that he or she has a personal responbility to use violence to settle disputes). Brutalization is not the same thing as child abuse (many groups endorse the violent domination of children and do not identify it as abusive), but child abuse can be and usually is brutalizing. So the best way to reduce youth violence is to prevent child brutalization and/or abuse, because children who are not brutalized never have to confront the choice that lead to violent criminality.

Programs designed to support medium-risk families with home visits by experienced mothers or nurses who counsel and coach have documented success at reducing injuries from abuse. So has the pioneering development of family community centers in some cities and states. A group of parents and local service providers in Addison County, Vermont, for example, started a parent-child community center in 1979 to offer area families services in the form of educational classes, support groups, child care, playgroups and recreation. The center also served as a focal point for coordinating the activities of state and local agencies concerned with children and family services and offered home visiting and school outreach. After seven years, teenage pregnancy rates in Addison County fell by 65 percent. The infant mortality rate was reduced to almost half the state average. Welfare dependency declined from 40 percent to 17 percent. Incidents of child abuse dropped from 21 percent to 2 percent. The Addison County center and several others like it proved so successful that Vermont has supported their replication in every county in the state.

Such family centers offer primary prevention to entire communities to help stabilize medium-risk families. Programs of heart-disease prevention similarly target the entire community rather than focus exclusively on heart-attack victims. High-risk families obviously need more intensive intervention.

But brutalization is only the first stage, the initiating trauma, in the four-stage process Athens identified in lengthy life-history interviews with violent criminals. Three further stages of violent social experience must follow to turn someone to violent criminality, which is why many brutalized children do not grow up to be violent adults. Intervention at any stage along the way can divert the novice from embracing violence, a decision which is based on a series of choices as well as on success with using violence first defensively and then offensively.

The second, belligerency stage of violentization reveals itself in threats, in an emerging cynicism and contempt (because the people and institutions which are supposed to protect the child from brutalization have failed him), in bullying and minor violent performances. The typical fate of belligerent students today in many jurisdictions is to be expelled from school, but expelling them from school simply throws them back into the maelstrom and cuts them off from help. Alternative high schools have been successful at least in part because they offer belligerent students alternatives, but Athens's evidence that boys usually complete violentization by the time they are fourteen years old means such help may come too late. We need alternative middle schools as well, or at least comparable middle-school programs, perhaps in a community center setting.

Violent coaching crucially influences a belligerent novice's decision to form his initial defensive violent resolution. This implies that nonviolent coaching should be an effective intervention provided that brutalization is not still ongoing.

The Menninger Clinic of Topeka, Kansas, is developing a Peaceful Schools Project which in pilot form has already documented dramatic reductions in school violence. The program, which includes defensive martial-arts training, zero tolerance for bullying and adult and high-school-student mentoring, evolved pragmatically and ad hoc but hews closely to Athens's violentization model.

Even at the third, violent performances stage, when the belligerent student has committed himself or herself to using serious violence defensively and has begun to do so, intervention is still possible. The violent performer still has choices to make - crucially, whether he will expand the range of his violence from defensive to offensive - from provoked violence to unprovoked. I suspect intervention at this crucial decision point accounts for the success of the Marine Corps in turning tough kids into responsible adults. The Marines teach recruits serious violence, but they constrain it to defensive violence within a code of honor and loyalty to the Corps. Something similar probably accounts for the success of programs that reach delinquents through training in the martial arts, which similarly invoke values of honor and defensive restraint. Athletic coaches in our schools, some of whom also function as violent coaches, would benefit from incorporating these distinctive values.

Once violentization is complete - once someone has committed a serious unprovoked or only minimally provoked violent criminal act - no one has yet found a reliable therapy or treatment to reverse it. But most violent crimes are committed by people between the ages of 15 and 30 - new graduates of the violentization process, so to speak - which implies that many violent individuals deescalate their violence as they grow older. Why and how they do so, and how they might be helped to make that choice sooner, is clearly a field ripe for further research now that Athens has demonstrated how they came to be violent in the first place.

But all the official programs in the world cannot replace personal witness to civil values; it is by personal witness, after all, that civil communities maintain their civility and the civilizing process proceeds that has gradually reduced personal violence in Western society. Athens's work discredits protestations that violence persists because of the poverty, race, culture or genetic inheritance of "those people over there" and has nothing to do with you and me. W. H. Auden once speculated that "Perhaps everything terrible is in its deepest being something helpless that wants help from us." Criminal violence emerges from social experience, most commonly brutal social experience visited upon vulnerable children, who suffer for our neglect of their welfare and return in vengeful wrath to plague us. If violence is a choice they make, and therefore their personal responsibility, as Athens demonstrates it is, our failure to protect them from having to confront such a choice is a choice we make, just as a disease epidemic would be implicitly our choice if we failed to provide vaccines and antibiotics. Such a choice - to tolerate the brutalization of children as we continue to do - is equally violent and equally evil, and we reap what we sow.