Guest Post by Anthony Cody

This post was originally published on Anthony Cody’s blog at Living in Dialog, on March 10.

In the mistaken belief that test scores are adequate reflections of learning, we have created vast systems to extract test score data, and now are requiring that this data be incorporated in teacher and principal evaluations across the country. Can we return to an authentic use of evidence and data in teacher evaluations?

The publication of teacher ratings generated by Value Added Models (VAM) in New York has prompted some closer examination of their validity, as I discussed here yesterday. This has brought into the mainstream what was, up until recently, a discussion among academics.

Even some of the proponents of these methods have quailed at the sight of teachers being scorned in the press for their low VAM ratings. Let’s assume that sanity prevails, and we are able to roll back these test-centered evaluation systems. Is there a way we SHOULD be incorporating student learning into teacher evaluations? YES, but we need to be very careful about how we do this, in order to avoid the pitfalls our reformers have led us into.

Student learning is at the heart of what teaching is all about. But we have made a huge mistake when we take what is a very indirect measurement of all the learning that ought to be occurring in a classroom — a test or set of tests — and mistake that for the sum of learning, and then start attaching negative and positive consequences to that. When we do this, instruction becomes distorted, and learning becomes all about test performance. The purpose for both teaching and learning is up-ended, and becomes about satisfying an external authority rather than pursuing what ought to be developed as an intrinsic passion. And this would be true even if VAM was accurate!

I became a National Board certified teacher more than a decade ago. This process revolves around providing compelling evidence that you, as a teacher, have helped students to learn. The National Board process uses as evidence student work over time, including teacher feedback, and the teacher’s analysis of how their instruction helped the student to develop. This is comparable to a portfolio that a teacher could assemble for evaluation purposes. It uses videotapes where you can see students engaging in scientific inquiry, analyzing data, and discussing their investigations. This is comparable to a teacher observation.

But part of what makes the National Board process worth pursuing is the richness of the National Board standards. They are NOT just about “student achievement,” which is a code word for test performance. They include other critical elements of effective teaching, like building classroom community, getting students actively engaged in rich dialogue, and responding to the needs of English learners and special ed students.

Many of these things are working against teachers who find themselves evaluated in the new VAM paradigm. As we are seeing in New York, teachers of Special Ed and English learner students are being ripped up in the press for their lousy results.

One year I found myself teaching several classes which included several students repeating 6th grade. I also had several students who appeared to me to be Special Ed, but their parents had made the decision to not have them classified in this way because they felt this would do more harm than good. Some of these students were highly disruptive, and the school system did not have adequate resources to respond to their needs. (Many of our schools have no nurses, no librarians, and minimal support staff.) These students affected the entire class.

We must have evaluation that is sensitive to the composition of each class. You will not get this from a spreadsheet. You cannot get here with data tools. We are dealing with human beings here. We need the skill and judgment of compassionate experts. And that is what we want every teacher to be with his or her students — a compassionate expert, able to give specific feedback, encouragement, and create a good learning environment.

A school is like a classroom in this way. A good principal is a strong instructional leader, and works with his or her staff as a community of learners. What we want is for the doors in that school to be open, for the principal to be circulating, for teachers to be circulating as well, observing and learning from one another, and solving problems together.

When the starting point for teacher evaluation reform is “we have far too many bad teachers, and we need to start using data to expose and drive them out,” the entire process is sabotaged from the start.

Data has a role to play, but the good uses for data are being de-legitimized by its systematic over-use. And the current atmosphere is driving teachers into a state of fear, which is absolutely destroying the trust and respect that is the pre-condition for effective feedback and growth.

I worked with a group of accomplished teachers a couple of years ago to produce a report that made substantive recommendations for improvements to teacher evaluation.

Here is a portrait of what I believe a good evaluation process ought to look like.

A teacher meets with his or her evaluator. They review the professional standards in use, and look for areas in need of growth. Maybe it is a focus on literacy and writing skills. Maybe it is bringing the English learners level of engagement and participation up. They discuss strategies the teacher might try to address these things, and they also discuss the forms of evidence they will look at over the year to see what is happening in this area. Assessment, especially of the classroom-based formative sort, is a powerful tool. How is a teacher assessing his or her students’ abilities? How are they using that information to give feedback and give the student appropriate, challenging work? This is where test data may play an important role, because a skilled teacher draws on this data to better understand their students.

Once an area of focus has been defined, the teacher and evaluator find some professional development resources that might help as well — maybe a conference to attend, some books that might be read, a grade level team that might come observe a lesson here and there and offer feedback, a colleague that is expert in this area to go observe. Then over the year, the teacher collects student work samples that provide evidence of learning. They document how they have designed instruction to help students learn, and show where they have provided feedback. The evaluator observes, a few times at random, and a few times by request, to see particular lessons. This evidence would be appropriate to the goal that has been set. It could include some test data, but test data would just be one source of evidence among many.

There is an idea put forth by some “reformers” that teachers are implacable opponents of any evaluation system that might reveal some of us to be deficient. Untrue. I worked for two years as a Peer Assistance and Review coach, and had the task of working directly with teachers who had received poor evaluations, to try to get them to improve their teaching. In my opinion, a process like the one I describe would be far more effective at “exposing” poor instruction than the VAM models we are attempting now. I did indeed work with some very ineffective teachers, and none of them could have “gamed” a system like the one I describe above, assuming a competent administrator were in place.

But the main benefit this approach offers is to reestablish the trust that is a pre-condition for effective collaboration, reflection and growth. When we have this kind of atmosphere, we can begin to honestly reflect on the evidence of how our students are growing and learning, and we can in turn, learn and grow as teachers.

Evidence of student learning does indeed have an important place in a good teacher evaluation process. But we must not center our evaluations on test scores, or on “student achievement,” or any other euphemism for test results. Even if we were able to create VAM systems that were more accurate than the current ones, this leads us into a dead end, where students and teachers alike are focused on answering whatever predetermined questions have been chosen to prove we have learned.

Learning must be dynamic and creative, and focused on challenges that engage our students. Their skills need to develop in the context of real-world projects, and these projects ought to reflect student interests and the communities in which they live. Great teachers track student learning closely, through a wide variety of assessment tools. A great teacher evaluation system ought to be able to do the same for our teachers, without making the mistake of equating test scores with learning.

What do you think? Should evidence of student learning be included in teacher evaluation? What should this look like?