Join the conversation

I think you skipped a step when you went to, "There are a number of performance-based evaluation systems that can determine whether teachers are meeting professional standards, such as peer review models."

The report specifies that value-added (and mostly based on one year of data which is a bit misleading) ALONE is not dependable for high-stakes decision making. That's not what these systems you attack are constructed to do. Most of the time value-added is but one measure amongst many, frequently including peer review or other improved observations as a major data source as well.

The true dichotomy is whether student test scores will be used or not-- peer review on professional standards = not and leaves that wide fissure behind. Using value-added not as a driver, but an informant is the compromise.

Given that VAM has high error rates, unstable ratings, false positives (the 5th grade teacher raising 4th grade scores), and insurmountable challenges trying to identify, let alone control for all of the factors that contribute to scores in such a small sample, I take no solace in the idea that some people would use it for smaller portions of evaluations. To acknowledge its shortcomings and then try to set an arbitrary percentage for its continued use in evaluations is illogical.

The valid use of state tests might be to look for evidence of what's happening in a school. Trends and abberrations in test scores might be an indication to look more closely at what's happening. But test scores are going to vacilate no matter what if you use enough of them over a long period of time. There's a lot of speculation involved if you read too deeply into them.

Anne makes an excellent point when she notes that faith in VAM results implies faith in the validity and usefulness of the tests used. At the moment, the very tests that were criticized for being low-bar and uneven, state to state, are now being positioned as "scientifically based" measurement tools through which we can identify bad teaching. Ironic.

I don't think anybody in this national discourse believes that annual testing will--or even should--go away. The core question is around how we use the test results. Assessment should inform instruction, serving as benchmarking for student progress and a guide for what's needed next. Using assessment data to punish or shame teachers is counterproductive.

VAM--a currently flawed methodology-in-development--might become a useful tool. But that implies expertise in translating results into a public understanding of complex statistical modeling. Policymakers do not understand the limitations of VAM. Journalists do not understand the limitations. Education leaders don't understand the limitations, either. Using VAM to evaluate and improve teacher performance is like using a jackhammer when you really need a scalpel.

This is a tiresome argument, and not one to help. This whole thing is based on acceptance of the teach, drill, test paradigm of the industrial model school. If we are to prepare students for this century, a change in perspective is needed. Assessment must take the form of classroom based formative student self-assessment with teacher supportive feedback on criteria and rubrics developed by students and teachers. Most people regard assessment as an entity unto itself, whereas it instead should be used as an important learning tool for and by the students. It is NOT best used to 'take the students' temperature' (or as a 'weed-out' device for teachers) but as a means for learning. Yes, it will show the students' progress, and for those who doubt the teacher's veracity, outside assessors could evaluate student work. And for those of you who think this won't work, we've had an action research group doing it since 2001 (on their own nickel - no grant!). The teachers have found higher achievement, intrinsic motivation, self-directed learning, higher order thinking, more productive habits of mind, and creativity in their students' work. Grades are determined by consensus of teacher and students, and the teachers say that the only point of contention is when students rate themselves lower on the rubrics than the teacher does. When the work is student-centered with strong student involvement, students are quite frank and honoest about their work. To 'bring this to scale' (new terminology!), a large public engagement would be required to convince politicians, the public, and especially educators that this is a necessary change in education. And the professional development necessary for educators would be huge. But it would be the way to prepare our students for the 21st century. However, it couldn't be done by the next election cycle, so politicians will dismiss it as airy visionary thinking instead of the tough stance they take in order to 'stand tall' and 'view with alarm' (or point with pride to the outliers that raise test scores ). (That makes me wonder, what would it be like if the Education Department had someone running it who actually was a teacher?)

I couldn't agree more with Mel. No evaluation system based on testing will work. The briefing paper also points out many negative consequences associated with using evalution systems tied to test scores, even VAM's. They include:

Discouraging teachers from wanting to work in school with the neediest students
Discouraging teachers from wanting to work with special needs students
Disincentives for collaboration
Narrowing of curriculum, group incentives tends to make narrowing even worse
Identifying wrong teachers for reward or discipline
Destroys morale
If it doesn’t accurately measure then how do you know how to improve?
Pushes out teachers
Won’t accomplish goals of identifying effective and ineffective teachers for reward and discipline.

I fear we are headed down this path with so much political momentum we can stop it. It is sad and it is a policy direction that will harm a whole generation of students.