There are good reasons to be troubled by the violence that pervades the media. Movies, television and video games are full of gunplay and bloodshed, and one might reasonably ask what's wrong with a society that presents videos of domestic violence as entertainment. Of course, the same questions could have been raised about watching Macbeth, or Punch and Judy. Let's face it, people have always enjoyed watching other people's mayhem.

Most researchers agree that the causes of real-world violence are complex. A 1993 study by the US National Academy of Sciences listed “biological, individual, family, peer, school, and community factors” as all playing their parts. And a 2001 report by the US surgeon general concluded that “the preponderance of evidence indicates that violent behavior seldom results from a single cause; rather, multiple factors converging over time contribute to such behavior”.

Viewing abnormally large amounts of violent television and video games may well contribute to violent behaviour in certain individuals. The trouble comes when researchers downplay uncertainties in their studies or overstate the case for causality. Sceptics were dismayed several years ago when a group of societies including the American Medical Association, the American Psychological Association and the American Academy of Pediatrics tried to end the debate by issuing a joint statement: “At this time, well over 1000 studies ... point overwhelmingly to a causal connection between media violence and aggressive behavior in some children.”

Freedom-of-speech advocates accused the societies of pandering to politicians, and even disputed the number of studies (most were review articles and essays, they said). When Jonathan Freedman, a social psychologist at the University of Toronto in Canada, reviewed the literature , he found only 200 or so studies of television-watching and aggression. And when he weeded out “the most doubtful measures of aggression”, only 28% supported a connection.

The critical point here is causality. The alarmists say they have proved that violent media cause aggression. But the assumptions behind their observations need to be examined. When labelling games as violent or non-violent, should Pac-Man gobbling a ghost really be counted as a violent event? And when experimenters measure physiological arousal, or record the time it takes game players to read 'aggressive' or 'non-aggressive' words from a list, can we be sure what they are actually measuring? The intent of the new Harvard Center on Media and Child Health, to collect and standardize studies of media violence in order to compare their methodologies, assumptions and conclusions, is an important step in the right direction.

Another appropriate step would be to tone down the crusading rhetoric until we know more. Several researchers write, speak and testify prolifically on the threat posed by violence in the media. That is, of course, their privilege. But when doing so, they often come out with statements that the matter has now been settled, drawing criticism from colleagues. In response, the alarmists accuse critics and news reporters of being duped by the entertainment industry. Such clashes help neither science nor society.