I love statistics in these situations, but they undermine you if they're the wrong ones. (I agree with your point though.)
You can't meaningfully say "you are more likely to die of domestic trouble than terrorists" in a general sense just before you go tempt them in a specific way.
The relevant statistic would be the number of theaters showing the film, divided by the number of theaters these people could reasonably attack, times the historical liklihood that similar threats are carried out, times a coefficient (either positive or negative) of local police effectivness (where they try to help or hinder the terrorists.)
Compare THAT to the liklihood of dieing to domestic troubles in the day immediately following a significant emotional event, selecting only from those sampled who have received actual threats from their loved one.