You're ignoring a lot of the factors that go into that improvement though. What if 5 of my 20 10th graders get into drugs? What if Dad gets back from the war and starts making Susie read? These type of things have an impact and they are beyond a teachers control. Not to mention you have a lot of problems with the making of tests. You're still relying on students taking high stakes test for the salary of someone else.
Well, if you are like most post elementary teachers, you have 6 classes of 25ish students per semester, for somewhere between 150-600 total students a year? Do you think that 5 students going on drugs, or one whose dad dies at war affects the aggregate average of all students in a statistically significant way? The math doesn't seem like it would. And beyond that, is the number of students going on drugs this year all that different than last year? Usually its not.
The idea that you can design a test that measures exactly how good the teacher is doing is not something I buy in the least bit. I see kids who are all A's in junior high turn into high school dropouts. We get warned of problem kids all the time, and sometimes they come into high school and flourish. It's just too much variance to accurately say "this teacher is doing a great job and this test shows it was all them."
Then how do you measure success? Saying "trust us as educations, we'll tell you how successful we are" is crazy, and nothing in the rest of the world works that way. If change in performance of your students isn't the metric, what is your goal as a teacher? just jam them into the next grade?
If designed properly I'd love merit pay, again I think that would drive my salary way up. I just don't think we know a foolproof way of doing it yet. Too much variance. I also don't see how you'd make it work for middle school and up, though it may be doable at the elementary level. It's much harder to judge my performance on merit when I'm one of 6 teachers a kid has. You can get a lot better gauge when kids have the one and same teacher all year.
Well, again, much of the time what is talked about is merit based bonuses, not direct merit based pay. Teachers would still keep their base salary, but teachers who excel would get paid higher.
I'll agree that designing something that is accurate would be difficult however teachers unions refusing to allow the conversation to begin, rather than them being a part of trying to put together a system that will work is outrageous.
The net gut feeling from parents is that the mediocre teachers are overpaid, and the good teacher are underpaid. My middle daughter had a kindergarten teacher last year who easily deserved twice her pay. It was a ton of factors, organizational skills, involving parents and students, communication, etc.
It as used in NYC for a while in 2007 I believe and failed miserably.
Some of the major reasons it doesnt work in education:
1. how do you judge success?
2. where does special ed fit in?
3. are tests scores the goal (if so, we see what happened with cheating scandals)
At the end of the day, it would be exceptionally difficult, even when using a portfolio based assessment as your gauge
Like i said above, success is based on improvement in skills in the same group of students. If you don't define success as that, what is your goal as a teacher? Special ed is no different, just the margins of improvement you expect are smaller.
If you can't test for the success, then you can't measure it, and its not real. With merit based bonuses, cheating would get you bonuses for a year, but then your baseline target improvement would be higher the next year