Student test scores should be used as one of multiple factors to measure a teacher's effectiveness, but they should not be the only measure. Test scores can indeed reflect the teacher's ability and thus need to be taken into account. However, there are situations, such as a teacher with a remedial class, where no matter how good the teacher, the students' scores are unlikely ever to match a class of more advanced students.
First, schools that are in wealthier areas tend to have better performing students than schools in poor, urban areas. So, all teachers are not starting from the same place, and judging them all by the same standards would not be fair. Second, this would encourage teachers to be more concerned with their students doing well on the test, rather truly understanding the subject matter.
I do not believe that they should be. I believe that there are some teachers that are good teachers and they help a majority of the students but there are always those few students that do not do well and there is really nothing that the teacher can do. This is not even based on the fact that we are constantly stuffing more and more kids into a room and it is getting to the point where a teacher is not going to be able to help all the kids if they do not want to learn.
No, I don't think that it is fair to use students' test grades in order to decide if a teacher is effective or not. All too often, students may not try hard on tests and just breeze through them not caring at all, which is something that teachers simply cannot control.
Test scores should not be the only measure of a teacher's effectiveness. Many students can learn from a truly good/effective teacher but may not test well. It seems that most "benchmark" tests these days are really only good measures of how well students test, rather than what they learn. A good teacher can still wind up with students who learn and retain the material but test poorly.