There's a difference between stomping on a defenceless harmless rabbit for the sake of an experiment and killing a lethal hostile threat in self defence.
I personally would refuse to kill the rabbit for an experiment to prove that I could kill it. I would not hesitate to kill someone threatening my family with harm.
I would kill the rabbit if I was to eat it. I have no philosophical or moral issue with eating meat and will kill and dress it with no problems (and have done). Besides, rabbit is tasty.
I do agree that this relucatance to kill has been noticed, most commonly with the military training new soldiers. However it's also been found that modern new recruits have less (or even no) restraints in being taught to kill (the theorised reasons for this are a whole can of worms that would probably derail the thread).
Except they don't multiply exponentially - only people who get bitten and get away successfully turn into zombies. The ones that don't get away, just get eaten (one of the reasons why a zombie apocalypse can't be taken seriously - their food source is also their form of reproduction, so if they're highly successful in 'hunting', there aren't many zombies; if they're not successful, they just get eliminated by the humans, assuming the humans aren't idiots and get their bitten treated or at least quarantined).
As for tightly confined space, yes and no. I agree to the small areas, I disagree to the confinement as all schools have fire exits and a known fire drill (and in Japan, they have regularly practiced earthquake drills as well).
In a school of ~1000 students, I doubt 12 infected could cause the entire school to be locked down, without being overrun and trampled by escaping students.
I agree that leaving the source and biology of the zombies as unknown works better, since it makes them a credible threat without breaking suspension of belief.