One of the most persistent myths in the culture wars today is that social science has proven “media violence” to cause adverse effects. The debate is over; the evidence is overwhelming, researchers, pundits and politicians frequently proclaim. Anyone who denies it might as well be arguing that the earth is flat.

Jonathan Freedman, professor of psychology at the University of Toronto, has been saying for almost twenty years that it just isn’t so. He is not alone in his opinion, but as a psychologist trained in experimental research, he is probably the most knowledgeable and qualified to express it. His new book, Media Violence and Its Effect on Aggression, surveys all of the empirical studies and experiments in this field, and finds that the majority do not support the hypothesis that violent content in TV and movies has a causal relationship to real violence in society. The book is required reading for anyone who wishes to understand this issue.

I should say at the outset that unlike Freedman, I doubt whether quantitative sociological or psychological experiments–useful as they are in many areas–can tell us much about the effects of something as broad and vague in concept as “media violence.” As a group of scholars put it recently in a case involving censorship of violent video games:

In a field as inherently complex and multi-faceted as human aggression, it is questionable whether quantitative studies of media effects can really provide a holistic or adequately nuanced description of the process by which some individuals become more aggressive than others.

Indeed, since “media violence” encompasses everything from cartoons, sports and news to horror movies, westerns, war documentaries and some of the greatest works of film art, it baffles me how researchers think that generalizations about “effects” can be made based on experiments using just one or a few examples of violent action.

Freedman, by contrast, believes that the experimental method is capable of measuring media effects. This may explain why he is so indignant about the widespread misrepresentations and distortions of the research data.

He explains in his preface that he became interested in this area by happenstance, and was surprised when he began reading the research to find that its results were quite the opposite of what is usually asserted. He began speaking and writing on the subject. In 1999 he was approached by the Motion Picture Association of America (MPAA) and asked to do a comprehensive review of all the research. He had not previously received organizational support and, as he says, “was a little nervous because I knew there was a danger that my work would be tainted by a connection with the MPAA.” He agreed only after making it clear that the MPAA “would have no input into the review, would see it only after it was complete, and except for editorial suggestions, would be forbidden to alter what I wrote. Of course,” he says,

they asked me to do the review, rather than someone else, because they knew my position and assumed or at least hoped that I would come to the same conclusion after a more comprehensive review. But there was no quid pro quo. Although I was nervous about being tainted, I am confident that I was not. In any case, the conclusions of this review are not different from those of my earlier review or those I expressed in papers and talks between 1984 and 1999.

The book proceeds meticulously to examine the approximately 200 studies and experiments that Freedman was able to find after an exhaustive search. (He suggests that the exaggerated numbers one often hears–1,000, 3,500 or simply “thousands” of studies–probably derive from a statement made by psychologist John Murray in the early 1980s when the National Institute of Mental Health sponsored a review of the media violence research. Murray said that there were about 2,500 publications of all kinds that were relevant to the review. This is far different, of course, from the number of empirical experiments and studies.)

Freedman begins with laboratory experiments, of which he found eighty-seven. Many commentators have noted the artificiality of these experiments, in which snippets of a violent film or TV show are shown to one group of viewers (sometimes children, sometimes adolescents or adults), while a control group is shown a nonviolent clip. Then their level of “aggression” is observed–or rather, something that the experimenters consider a proxy for aggression, such as children hitting a Bobo doll (an inflatable plastic clown), delivering a “white noise” blast or–amazingly–answering yes when asked whether they would pop a balloon if given the opportunity.

As Freedman and others have pointed out, these laboratory proxies for aggression are not the real thing, and aggressive play is very different from real-world violent or destructive behavior. He comments:

Quite a few studies with children defined aggression as hitting or kicking a Bobo doll or some other equivalent toy…. As anyone who has owned one knows, Bobo dolls are designed to be hit. When you hit a Bobo doll, it falls down and then bounces back up. You are supposed to hit it and it is supposed to fall down and then bounce back up. There is little reason to have a Bobo doll if you do not hit it. Calling punching a Bobo doll aggressive is like calling kicking a football aggressive. Bobos are meant to be punched; footballs are meant to be kicked. No harm is intended and none is done…. It is difficult to understand why anyone would think this is a measure of aggression.

Freedman notes other serious problems with the design of lab experiments to test media effects. When positive results are found, they may be due simply to the arousal effect of high-action entertainment, or to a desire to do what the subjects think the experimenter wants. He points out that experimenters generally haven’t made efforts to assure that the violent and nonviolent clips that they show are equivalent in other respects. That is, if the nonviolent clip is less arousing, then any difference in “aggression” afterward is probably due to arousal, not imitation. Freedman’s favorite example is an experiment in which one group of subjects saw a bloody prizefight, while the control group was shown a soporific film about canal boats.

But the most striking point is that even given the questionable validity of lab experiments in measuring real-world media effects, the majority of experiments have not had positive results. After detailed analysis of the numbers that the researchers reported, Freedman summarizes: Thirty-seven percent of the experiments supported the hypothesis that media violence causes real-world violence or aggression, 22 percent had mixed results and 41 percent did not support the hypothesis. After he factored out experiments using “the most doubtful measures of aggression” (popping balloons and so forth), only 28 percent of the results were supportive, 16 percent were mixed and 55 percent were nonsupportive of the “causal hypothesis.”

For field experiments–designed to more closely approximate real-world conditions–the percentage of negative results was higher: “Only three of the ten studies obtained even slightly supportive results, and two of those used inappropriate statistics while the third did not have a measure of behavior.” Freedman comments that even this weak showing “gives a more favorable picture than is justified,” for “several of the studies that failed to find effects actually consisted of many separate studies.” Counting the results of these separate studies, “three field experiments found some support, and twenty did not.”

Now, the whole point of the scientific method is that experiments can be replicated, and if the hypothesis is correct, they will produce the same result. A minority of positive results are meaningless if they don’t show up consistently. As Freedman exhaustively shows, believers in the causal hypothesis have badly misrepresented the overall results of both lab and field experiments.

They have also ignored clearly nonsupportive results, or twisted them to suit their purposes. Freedman describes one field experiment with numerous measures of aggression, all of which failed to support the causal hypothesis. Not satisfied with these results, the researchers “conducted a complex internal analysis” by dividing the children into “initially high in aggression” and “initially low in aggression” categories. The initially low-aggression group became somewhat more aggressive, no matter which programs they watched, while the initially high-aggression group became somewhat less aggressive, no matter which programs they watched. But the children who were categorized as initially high in aggression and were shown violent programs “decreased less in aggressiveness” than initially high-aggression children who watched neutral programs. The researchers seized upon this one highly massaged and obscure finding to claim that their results supported the causal hypothesis.

Freedman examines other types of studies: surveys that compare cities or countries before and after introduction of television; experiments attempting to assess whether media violence causes “desensitization”; longitudinal studies that measure correlations between aggressiveness and preference for violent television over time. No matter what the type of study or experiment, the results overall are negative. Contrary to popular belief, there is no scientific support for the notion that media violence causes adverse effects.

Why, then, have not only researchers and politicians but major professional associations like the American Academy of Pediatrics and the American Medical Association repeatedly announced that thousands of studies have established adverse effects of media violence? One reason was suggested to me recently by a pediatrician active in the AAP. The organization’s guidelines argue for scientific support for policy statements. This puts the AAP in a serious bind when, as is the case with media violence, its leaders have a strong opinion on the subject. It’s tempting then to accept and repeat assertions about the data from leading researchers in the field–even when it is distorted or erroneous–and that’s what the professional associations have done.

Another factor was candidly suggested by Dr. Edward Hill, chair of the AMA board, at a panel discussion held by the Freedom Forum in New York City last year. The AMA had “political reasons,” Dr. Hill said, for signing on to a recent statement by professional organizations asserting that science shows media violence to be harmful. The AMA is “sometimes used by the politicians,” he explained. “We try to balance that because we try to use them also.”

Because Jonathan Freedman believes the scientific method is capable of measuring the impact of media violence, the fact that it hasn’t done so is to him strong evidence that adverse effects don’t exist. I’m not so sure. I don’t think we need science to know from observation that media messages over time can have a powerful impact–in combination with many other factors in a person’s life. Some violent entertainment probably does increase aggression for some viewers, though for as many or perhaps more, the effect may be relaxing or cathartic.

If the media do have strong effects, why does it matter whether the scientific research has been misrepresented? In part, it’s precisely because those effects vary. Even psychologists who believe that the scientific method is relevant to this issue acknowledge that style and context count. Some feel cartoons that make violence amusing have the worst effects; others focus on stories in which the hero is rewarded for using violence, even if defensively.

But equally important, the continuing claims that media violence has proven adverse effects enables politicians to obscure known causes of violence, such as poverty and poor education, which they seem largely unwilling to address. Meanwhile, they distract the public with periodic displays of sanctimonious indignation at the entertainment industry, and predictable, largely symbolic demands for industry “self-regulation.” The result is political paralysis, and an educational structure that actually does little to help youngsters cope with the onslaught of mass media that surround them.