I know (believe) there is a literature on the relation between violent media content (movies, TV, video games) and behavior (perhaps in children, but maybe more broadly). ...[I do not know] if it uses credible methods. [But] if any of it does, could it at least suggest that media can make an impression that makes certain actions more likely?
This is an interesting and difficult question. There are well-conducted studies -- notably by the social psychologist Brad Bushman and his colleagues -- that show that, for example, exposing a child to a violent videogame, leads to an increased likelihood of aggressive behaviors. Craig Anderson, Bushman, and their colleagues conducted a meta-analysis of this research and concluded that
The evidence strongly suggests that exposure to violent video games is a causal risk factor for increased aggressive behavior, aggressive cognition, and aggressive affect and for decreased empathy and prosocial behavior.
Which suggests that if we had a large increase in the consumption of violent videogames, we'd see a large increase in violence, right? However, we don't seem to see this. The Figure to the right plots the rates of assaults in the US as reported in the Bureau of Justice Statistics Victimization Survey. When the survey started, there were no such things as violent videogames. But the rates of assaults have steadily decreased. Just for fun, I put in the date when the first version of Grand Theft Auto was released. If anything, the decline in assaults seems to have accelerated during the violent videogame era.
Similarly, look at the rates of rapes reported in the survey. Again, there has been a steady decline. Notice that the decline was not affected by the appearance of the world wide web. I don't have data, but it seems clear that the web has greatly facilitated access to pornography, and that exposure of contemporary adolescents is orders of magnitude greater than was true for my cohort. Nevertheless, there is a lot less rape.
So, does that mean the experimental social psychology studies have been done wrong, and that media violence has no effect? No, and No. It is entirely possible that if we hadn't had violent videogames during the last 15 years, the assault rates would have dropped even further. That said, the Figures above suggest that the effects of violent videogames on actual violence may be small relative to other social forces that have reduced rates of assault and rape, and similarly for the effects of pornography on sexual aggression.
By the way, if you have read this far, stop and give thanks for the improvement in our quality of life that is reported in these Figures.
So what do we make of this? It's plausible to me that, as Austin suggests, the effect of violent rhetoric is something like the effect of violent videogames or pornography. The data suggest, then, that if media exposure to violent political rhetoric has an effect, it's likely to be harmful, but if so the harm may be small.
So am I arguing that violent political rhetoric is okay? Well, would you encourage children to play Grand Theft Auto? Would you direct an adolescent to a violent pornographic website? I wouldn't, even if I thought that the chance of lasting harm was small.