Well, I really mean to answer “what’s wrong with using the results of students’ standardized tests to evaluate teacher performance?”
On its surface, nothing. Indeed, student performance should be the primary metric of teacher quality. When I worked for Bodsat Prep, I was gauged by that metric myself: what matters is the point gain.
The main problem, though, is that the standardized tests don’t test everything that we want our students to know. In the same way that the standardized test at the DMV doesn’t tell you whether you’re a good driver (though it does catch some really bad drivers), standardized testing for schools doesn’t tell you whether you’re on track to be a productive member of society (though it does catch a number of people who really aren’t).
Also, if a standardized test samples say a random 10% of what one should know, then how well one does on such a test is, statistically, a good indicator of how much of the target material one knows. But if it’s a non-random 10%, then you can count on prepared students to know that 10% very well. And possibly nothing else.