TCS Daily

TV in the Dock

By Iain Murray - March 17, 2003 12:00 AM

Almost exactly a year after the last attempt, here comes another study aimed at showing America that its television viewing habits cause violence. Researchers from the University of Michigan have found that children who prefer violent TV shows are more likely to be violent adults. This research adds to the vast literature indicting TV as a cause of violence. Yet once again, the case must be dismissed for lack of evidence.

The study (currently available here) was a long time in the making. In the late 70s, the researchers interviewed the subjects as children, asking them if they which of 80 popular TV shows they watched, if they identified with violent characters and how realistic they thought the shows were. They categorized the shows by violent content, enabling them to score the children by their liking for violent TV shows.

This is where the first objection to this study arises. Roadrunner cartoons were characterized as "very violent." This might well strike the average TV viewer as a rather broad definition of violence, especially given that virtually everything that happens to Wile E. Coyote in those shows is either accidental or self-inflicted. I am unable to think of an occasion when the roadrunner employed any method more violent than delivering a surprising "Meep meep" while the hapless cartoon carnivore's back was turned. It might therefore be useful to see what the results would be like if a more traditional definition of a violent show were to be used.

The researchers found that children who scored highly in response to any of the questions were more likely to be violent or law-flouting in later life. This held true for both sexes and regardless of social status, intelligence, parents' child-rearing practices or even childhood aggression. The researchers concluded that there was a causal link between violent TV and violent behavior later in life and recommended parents restrict their children's viewing of violent programs and movies.

Yet this may be a step too far. The associations the researchers found were very weak. No correlation for actual physical violence was stronger than the 95 percent confidence level. This means that the chances of TV not being the cause are about 1 in 20, but this definition of statistical significance has come under increasing criticism, with the British Medical Journal recently suggesting that a confidence level of 1 in 100 or even 1 in 1000 should be required to demonstrate significance. The correlation between girls who liked violent shows and "indirect aggression" was stronger at the 99 percent level, but this was not the case among boys.

The weakness of the associations can be demonstrated with a few examples. "High-violence watching" boys were more likely (again, only at the 95 percent confidence level) to have shoved a person as an adult than other boys, but were no more likely to have punched, beaten or choked another adult. They were no more likely to report having committed a crime or a traffic violation, but were more likely to have had a state recorded crime or traffic violation (although this is based on a very small number of incidences). The small size of the study (153 boys and 176 girls from two areas of Illinois) also calls into question how readily the findings can be generalized.

Yet the researchers developed their own model on top of these weak associations called "Adult Composite Aggression," combining all sorts of aggression, whether it be physical, indirect or traffic violation. When they compared TV viewing to this aggregate measure, they did find correlations at the 99 percent confidence level or better. The question must then arise how accurate the researchers' model is. The lack of strong associations between TV viewing and more traditional definitions of violent behavior suggests that the model may overstate the case. The researchers may have found a strong link between TV violence and general "aggression," broadly defined in the sense that a driver can be aggressive, but they did not really find a link to violence per se.

It is also interesting that studies that find a link between TV violence and aggressive behavior in some ways contradict other studies that find a link between TV in general and heightened aggression. The former say that children watching "Sesame Street" is fine, as long as they don't watch "WWE Smackdown." The latter suggest that any form of television viewing among children is bad, and that Cookie Monster will lead eventually to aggravated assault. Can both be true? Possibly, if TV increases aggression and violent TV increases it further.

But if that is the case, then we should be seeing it show up in the crime figures. The early 90s were a time of increasing violence on TV, with shows like "Walker: Texas Ranger" providing numerous violent fights each episode. This was a far worse era for TV violence than that of Roadrunner and The Six Million Dollar Man. Yet the children whose minds may have been poisoned by these offerings are noticeably less violent than their predecessors. The juvenile homicide rate has been falling steadily since 1993, and is much less than it was when the roadrunner generation were in their mid-late teens.

The link between TV and violence is overstated. There may indeed be a link between violent TV and general overconfidence, arrogance and confrontational behavior, but that doesn't seem to translate into increases in crimes and real, physical violence. TV, including violent TV, gives us a great deal of enjoyment, which is why we watch it so much. It doesn't seem as if the social cost of that enjoyment is one we mind paying.

TCS Daily Archives