Moore's resolution (Americans watch more frightening TV news) seems questionable.
First, Canadians watch a lot of American news -- one estimate is that they watch more American news than Canadian (60:40).
And early in the movie he discards the possibility that playing violent video games and watching violent flicks can cause violence -- because Canadians like, and Japanese positively love, those.
If violent movies and violent videogames cannot cause violence -- then how can newscasts about violence do so?
One can see how newscasts could instill the fear he talks about. But linking that manner of fear to homicide is another matter.
Homicides fall into two major classes. There's the killing in the course of another felony -- crack dealers removing the competition, armed robbers who kill their victim. Can't see much link to media-induced fears there.
The other and larger group is crimes of passion. A couple of people get very drunk, fly into an argument, and one kills the other. Again, hard to see how this links to fear induced by watching the news. I suppose fear might make people jittery, and jittery people are easier to provoke, but it seems hard to explain major homicide rate differences by this means.
Of course, it's hard to find any good theory to explain variations in homicide rates. US gun homicide rates fell by about 40% between 1993 and 1999, although I don't recall the news getting any less messy, about 3 million new firearms a year entered the market, and economic disparities between rich and poor increased somewhat.