Grassroots Commentary

Childhood TV Watching and Crime

By John Ray · Feb. 22, 2013

The study outlined below has excited both liberals and conservatives but it has been much overhyped. See my comments following the report:

Children who watch excessive amounts of television are more likely to have criminal convictions and show aggressive personality traits as adults, a New Zealand study has found.

The University of Otago study tracked the viewing habits of about 1,000 children born in the early 1970s from when they were aged five to 15, then followed up when the subjects were 26 years old to assess potential impacts.

The research, published in the US journal “Pediatrics” this week, found a strong correlation between childhood exposure to television and anti-social behaviour in young adults.

“The risk of having a criminal conviction by early adulthood increased by about 30 percent with every hour that children spent watching television on an average weeknight,” co-author Bob Hancox said.

The study also found excessive TV viewing was linked to aggressive personality traits and an increased tendency to experience negative emotions.

It said the links remained statistically significant even when issues such as intelligence, social status and parental control were factored in.

“While we’re not saying that television causes all anti-social behaviour, our findings do suggest that reducing television viewing could go some way towards reducing rates of anti-social behaviour in society,” Hancox said.

He said the findings supported the American Academy of Pediatrics' recommendation that children should watch no more than one to two hours of quality television programming a day.

The study said it was possible that children learned anti-social behaviour by watching it on TV, leading to emotional desensitisation and the development of aggressive behaviour.

But it said the content of what children were viewing was not the only factor, highlighting the social isolation experienced by those who spent hours watching the box.

“It is plausible that excessive television viewing contributes to anti-social behaviour in ways unrelated to violent content,” it said.

“These mechanisms could include reduced social interaction with peers and parents, poorer educational achievement, and increased risk of unemployment.”

Hancox said the study concentrated on children’s viewing habits in the late 1970s and early 1980s, before the advent of personal computers, and further research was warranted into how such technology affected subsequent behaviour.

My immediate reaction to this report was that we are just seeing another social class effect here: Working class children are more likely to come from criminally-inclined homes and also are more often left to be “minded” by the TV.

Such effects are well known, however, so the authors were unusually conscientious and controlled for them. They used analysis of covariance to remove the effect of class variables. And what did they find when they did that? I quote: “After controlling for additional covariates, associations between viewing time and criminal conviction and antisocial personality disorder remained statistically significant, although the association between television viewing and violent convictions did not.”

So I was pretty right. Watching a lot of TV as a kid does not of itself make you more likely to be a violent criminal but coming from a lower social class does. It is only non-violent criminality (presumably drug offenses and the like) that is somewhat associated with childhood TV viewing.

The study is actually good evidence AGAINST the concerns of the TV haters. Reading the actual “Results” section of academic journal articles has long been a pesky habit of mine. Sorry to puncture any treasured bubbles

AND MOST IMPORTANT: ALL of the correlations reported were trivially small. They were statistically significant only by virtue of the large sample size. By any criterion of real-life significance, TV viewing predicted NOTHING. To put that another way, TV viewing explained at best only around 2% of anything else