I think you can evaluate teachers by evaluating them against their peers (same grade level) within their school district.
Here in Connecticut, we have standardized tests (Connecticut Mastery Tests, or CMTs) that every child in the state in grades 3 through 8 take every year. Scores are summarized both statewide and district wide. If some teachers consistently have classes whose aggregate CMT scores are averaging below other teachers at the same grade level within the same school, or within the same school district, then that should be a clue about ineffectiveness.
I'm guessing that other states have similar standardized tests that can use this same approach to evaluating teachers.