Or are you simply of the Nicholas Garaufis school of thought; namely, if the test isn't perfect, it can't possibly be used as a hiring tool? After seeing what Judge G did to the FDNY training program, I'd have half a mind to buy extra smoke alarms and fire extinguishers if I lived in the city...
Like a lot of people, I know some persons who would probably be excluded by the general approache I described upthread, and who ended up going into teaching and being really, really good at their specialty. That's why there should always be a way to make exceptions. That doesn't invalidate the principle that "book smart" isn't a bad criteria for new teachers, and that some tests can give a rough idea of how likely someone is to be so.
Cliched but true:
Tough cases make bad law. The plural of anecdote isn't data.