Violent Games and Crime Correlations

Saw this on Digg.com, and it reminded me of something I recently read in Anderson, et al.'s book on violent video game effects.

"The old Logic 101 principles regarding the establishment of a factor as being a necessary and sufficient cause of an effect simply don't apply to most modern science (Anderson & Bushman, 2002c). We know that smoking tobacco causes an increase in the likelihood that one will contract lung cancer, but not everyone who smokes gets cancer, and some who don't smoke do get lung cancer. The probabilistic nature of modern science is largely due to the fact that multiple causal factors are involved in most medical, psychological, and behavioral phenomena. And for this reason, the old necessary and sufficient rules simply do not apply. Thus, every time people argue that violent video games can't be considered causes of aggression because they have played such games and haven't killed anyone is committing a major reasoning error, applying the 'sufficient' rule to a multiple cause phenomena. A similarly invalid argument is that the reduction in U.S. homicide rates during the 1990s - while violent video games were becoming more prevalent - proves that violent video games can't cause increases in aggression. This argument assumes either that violence is not caused by multiple factors, or that those factors are unchanging over time. Neither assumption is true (consider, for example, changes in overall incarceration rates, federal gun registration laws, drug use patterns, age distribution of the population, poverty rates, employment patterns, war), rendering the argument so weak as to be embarassing.

Though I did get tired of how combative and dismissive the Anderson, et al. book is in places, I think this is a very good point made. Most of Anderson's work has been in the laboratory, which allowed him the freedom to isolate the media exposure mechanism from the numerous other factors that cause aggression in young people. Moving such a project from the lab to the real world is the next step, but his point about multi-causality and the need to separate correlation from causality cannot be overstated. Tell that to the 1,541 others who have voted to digg this picture, though.