I'm skeptical about that, but for different reasons. Yes, it's potentially cheap (though could be expensive if it resulted in school district liability in wrongful death claims associated with those weapons). But I think the risk of accidental shootings would exceed the risk you're trying to address.
We're talking about around 50 deaths per year we're trying to prevent. Let's say you could cut that in half if you had an average of three pistol-packing teachers per school. Great. 25 lives saved. Meanwhile, you have 300,000 pistols introduced into school settings. How frequently will one of those teachers suffer some sort of mental break, with the presence of a nearby gun making that break a deadly occurrence? How frequently will a gun go off accidentally, in a bag, and kill someone? How frequently will a teacher gun down someone in a panic who was, in fact, harmless (e.g., a kid playing around with a real-looking toy gun? Or even a kid brandishing a real gun who would not have fired?) How frequently will a teacher responding to a real event manage to kill an innocent bystander with a misplaced shot? How frequently will a troubled kid get his hands on the teacher's gun by going through her desk or purse, and then shoot someone? I don't know the answer to those questions. However, if I ascribe even infinitesimal probabilities to each such event, and multiply that over 300,000 guns, I think we'd very easily get more shooting deaths than we'd be preventing.