joyce, i think there are two issues here. one is: can these people produce decent m-level research? and the other is: will this attempt at doing research make them better teachers (or make their learning experience more interesting)?
as to the first, i would say it's not impossible but would depend on many factors and as such could not be expected; because, as you say, normally before we do research or something, to find out something new, we study what is known already first.
as to the second, well, what better way to learn about the strength and weaknesses of different teaching methods than by trying them out, experimenting with them? if someone tells me: you've got to teach like THIS. i can learn that by rote, but i won't really have understood it. but if i try to improve that (and perhaps spectacularly fail) then i will have truly understood why this is how it is done!