I want to thank Jim Manzi for writing a truly excellent post on teacher evaluations. His points 1 and 4 in particular are almost universally not understood.
I do want to add a few additional points of my own:
1. Evaluations establish the principle that there is such a thing as performance in the first place. A great deal of discussion nowadays in education revolves around the idea that what we need to “fix the schools” is great teachers. But if that’s what we need, we’ll never do it. What we need, instead, are mechanisms for getting marginally better performance, year after year, from a teaching pool that remains merely adequate.
One bit of low-hanging fruit for achieving that goal, meanwhile, is the ability to dismiss the bottom 5% of teachers in terms of performance. Not only are these teachers failing comprehensively in their own classrooms, but their mere presence has a corrosive effect on an entire organization – on the teachers, on the students, on the management of the school. But right now, firing these teachers is essentially impossible. For all the difficulty of doing a rigorous evaluation in order to improve teaching performance across the board, I suspect it is a whole lot easier to identify the worst teachers in the school. If that could be done, the pressure to be able to terminate them would be significant, and that could do a lot to improve school performance right there.
2. Value-added metrics wind up punishing perfectly good but not spectacular schools with above-average student bodies. It may be that these schools should suffer reputationally, because the staff is not actually delivering as much value as they should. But high-stakes standardized testing actually pushes these schools to destroy themselves, wiping out the programs that actually do deliver value to these high-aptitude students and instead focusing on teaching to the tests.
That’s not an argument against using value-added metrics as such. It’s an argument that they need to be used intelligently, with some understanding of what “value-added” means at different points on the performance spectrum. But that, in turn, would require admitting that different standards are needed for students with different aptitude, which, in turn, is extremely difficult for our education system to admit. (And, admittedly, it’s a problem in corporate cultures that cross widely different customer bases as well. How well would Wal-Mart manage Tiffany?)
3. Nobody goes into teaching “for the money” – that is to say, teachers in aggregate make significantly less than people with their educational credentials and academic aptitude could make in other professions. So monetary rewards are useful primarily going to prove useful as signaling devices. There’s a lot of evidence coming in from high-performance charter schools suggesting that a monetary reward system tied too closely to evaluations actually degrades performance, because it gets teachers focused on the evaluations rather than on the performance. The evaluations should primarily be used as a diagnostic, to identify correctable deficiencies in teacher performance so they can be corrected through staff development, and to identify gross deficiencies in teacher performance so the teachers in question can be dismissed.
4. Similarly, across a system, what evaluations are useful is for research purposes and to drive market discipline. Evaluations of a school should be very useful to parents seeking to select a school for their child. Schools that consistently achieve high valuations (particularly for value-added metrics) should be objects of study by administrators and others looking to replicate that performance in lower-performing but still basically well-run schools. The least-important use of the evaluation is to directly “reward” or “punish” a school bureaucratically – and, indeed, if that becomes the primary use then the school is likely to start focusing overwhelmingly on the evaluation process and lose sight of actual performance. I’ve seen this happen over and over in New York City schools; it’s not a theoretical question.
As with almost all social reforms, the cause of school reform is best served by under-claiming what any particular reform can achieve. Try to move the needle a little bit with each change in the right direction, and try to line up incentives properly. Wild misrepresentations of what can be done, like the movie Waiting for Superman, actually undermine the very case they are trying to make.