05/12/2015 08:27 am ET Updated May 12, 2016

Do Genetics Explain the (Sometimes Seen) Relationship Between Television Viewing and Antisocial Behavior?

Whether exposure to television violence contributes to antisocial behavior has been an issue of significant controversy for almost five decades. Some studies find evidence for effects, other studies don't, and there is much debate among scholars regarding the meaningfulness and quality of the studies on both sides. Some scholars have been willing to argue that as much as half of all homicides in the US might be eliminated if childhood exposure to violent television (and other media) were reduced. This debate has also been the source of some "scientific urban legends" such as the old claim, repeated by the American Academy of Pediatrics (AAP), that 3500 studies of media effects existed by the year 2000 with only18 not finding evidence for harm (it turns out these numbers were apocryphal and don't remotely represent the actual state of the field). Put simply, it's a debate where there's a lot of bad information out there, sometimes promulgated by folks who ought to know better.

One of the ongoing questions in this debate is, if small correlations exist between television viewing and antisocial behavior, is this a causal effect of too much television, or are other factors at play? Do more aggressive people seek out more violent television, for instance? On a very simple level boys tend to watch more violent television and are also more aggressive than girls, so any observed correlation may simply be a gender effect. Unfortunately not all studies include appropriate controls (be particularly wary when you read about meta-analyses which routinely report results without any proper controls.)

One study that got attention back in 2013 (and released by the ever present AAP) was a longitudinal study from New Zealand (henceforth the Kiwi Study), which presented data suggesting that early television viewing is associated with antisocial behavior in adulthood. The Kiwi study examined overall television viewing, not violent television specifically, but since many media watchdog groups assert that television is soaked in violence, this is probably a fine proxy. The authors deserve credit for also controlling for many other relevant variables such as gender, IQ, parenting environment, etc. Doing so predictably weakened their results...associations between television and violent convictions became non-significant, and associations for any convictions and antisocial personality disorder became small, but remained significant. So for at least two out of three outcomes, television is associated with antisocial traits, albeit weakly. Despite that their data is correlational, the authors purport 'The findings are consistent with a causal association and support the American Academy of Pediatrics recommendation that children should watch no more than 1 to 2 hours of television each day.' That the AAP arguably has a conflict of interest in publishing research cheerleading their own policy positions is a whole separate issue I hope to get into in a later post. But asserting causality from correlational data is a basic no-no (I also chuckle a bit at a typo in the abstract suggesting some of the people in the study were born in 1773, but I'm typo prone enough myself not to get too much hubris over it).

However, a more recent study suggests that the Kiwi Study and others like it failed to control for at least one key variable: genetics. Published by criminologists Joseph Schwartz and Kevin Beaver, this new study is very similar to the Kiwi Study in that it follows a group of youth over time to examine the impact of television viewing in childhood on adult antisocial outcomes. Like the Kiwi Study, the authors control for multiple other variables. Doing so greatly reduces the correlation between television viewing and adult antisocial behavior, but does not eliminate it. Different from the Kiwi Study, Drs. Schwartz and Beaver had access to sibling pairs including twins, allowing for an examination of genetic effects. With genetic effects included, associations between television viewing in childhood and adult antisocial behavior vanished. The authors argue that genetic development is the key causal variable linking both a tendency to view more violent television and engage in later antisocial behavior.

Controlling for genetics is generally difficult and expensive to do in most studies. But without such data we have to remember we're not seeing the full picture. We also have to remember that the effects we're discussing are very tiny (not, despite more apocryphal claims by the AAP, anywhere similar to important medical outcomes), a fact the scholarly community has been remiss not to be more honest about with the general public. For half a century now, the general public has been hammered over and over by groups like the AAP with hyperbole that television viewing is harmful. The study by Drs. Schwartz and Beaver is just the latest in an increasing array of studies to suggest directly linking media use to antisocial outcomes was a mistake. As the data increasingly turns against causal effects, groups like the AAP will find themselves increasingly out on a limb, given how much of their reputation they have invested in needlessly frightening parents for decades. How they respond to these changes in the zeitgeist in the field and the data coming out will be interesting to watch. It may be as dramatic as a night of television!