The problem with "research" is it gets boiled down to statistics. Almost no one looks at the experiment design. Even if the statistics part is done well it is very easy to find a misleading correlation most people will assume is causation.
It is very easy to say that the shooters all played violent video games 3 hours a day or watched violent movies and that CAUSED the violence. In reality it is more likely they had a drive for violence/empowerment and were using video games and movies as a release/expression. The "research" will almost certainly conclude the former though.
The same with firearms and everything else.