Thinking about teacher evaluation

Teachers should not be subjected to the corporate numbers game advocated by StudentsFirst and the corporate education reformers.  But, still, we do know that some teachers are better than others.  And in a rough way we know how to identify, say, the better 4th grade teachers: ask the 5th grade teachers.

That’s not an evaluation system, though.  It’s too subjective.  And there is a lot to consider.  The kids have different backgrounds and levels of readiness to learn.  They have different styles of learning.  Also,  many teachers may have helped any one student, so we don’t always know who is responsible for a student’s progress.

So the question is how to distinguish the better teachers from those who need improvement.  The number of years a teacher has taught and the number of courses taken probably isn’t it.  Classroom observation is a component of it – if the observer is very good at it.  But there’s got to be more to it than that.

Corporate-stye education reform

That’s why education reformers propose measuring teacher performance based on student test scores.  The trouble is, when deployed by radical advocates of corporate-style education reform,high stakes testing can become a club for beating up teachers and their unions, shutting down schools and privatizing public education.

You see the corporate style in the recently broadcast Frontline documentary by John Merrow, “The Education of Michelle Rhee.”  You can go to minute 4:55 of the broadcast (here) and watch Michelle Rhee taking the classic corporate approach: “Give me improved proficiency (profit) numbers or you’re gone.  Don’t tell me how.  Just do it.”  During Ms. Rhee’s 3 year tenure in D.C., you never heard discussion of new instructional approaches, curriculum strategy or getting teachers the textbooks they needed.  And if the teacher were to point out that the kids are starting way behind and have tough family situations, it would just sound like excuses.  These are all the recognizable features of corporate-style education reform.

The logical outcome of the corporate style of school management is teaching to the test.  And here (listen from minute 7:00 to minute 8:30) you see it in action.  A teacher is doing just what you would expect.  She is not talking about how exciting the upcoming math and literature lessons will be but is trying to get the students’ buy-in to a flawed national education policy (and teacher job security).  Then an administrator explains that Michelle Rhee made those test scores the “be all and end all” for measuring principal performance.

And if it all works, the district superintendent can point to improved “performance,” as Ms. Rhee does.  Just as in a corporate setting, that short term performance may indicate a good product or just a manipulation of the numbers.  When companies cheat, the shareholders might lose, but when our schools teach to the test, our kids enter the world unprepared.

UPDATE 2/11/13: John Merrow,  in a recent report on his blog, provides further insight into the results Ms. Rhee achieved with her corporate-style  management of the Washington D.C school system.

The New Hampshire approach: evaluation as coaching

But there’s a big difference between the corporate evaluation model aimed at firing employees who don’t produce “the numbers,” and the coaching evaluation model, aimed at helping the faculties and schools improve.  Educators do not have it all figured out.  Measuring kids’ progress and attributing that progress to specific teachers is hard to do right.  The discipline is in its infancy.  But it does not have to be unfair and destructive as the headline-grabbing, union-busting corporate model is.

The New Hampshire Department of Education’s “Educator Support and Evaluation System” promises to be a step in the right direction.  Here is the draft that is currently circulating.  The final version should be out any day.

It’s long, but here is a summary.

Right away you see that the New Hampshire model looks nothing like the corporate education reform model.  As a first step, a task force assembled by the New Hampshire DOE identified standards for teacher performance based on an approach widely-used in the US.

To find out how well a teacher meets those standards, evaluators will look at:

• a teacher’s own assessment of her work and her goals for improvement,
• her lesson plans and other preparation materials, and
• classroom observations and discussions with the teacher

And testing is a piece of it

You could say that all that is about the process of teaching and how good the teacher appears to be at it.  The final component of the evaluation, roughly 20% of a teacher’s score, is the results visible in student performance – how well the teacher’s students have learned what they were supposed to learn.

This is where value added modeling or other way of attributing student achievement to teachers comes in.  There are many problems with the established ways of using test scores to evaluate teachers but the New Hampshire proposal is meant to be a major improvement.  Like some 15 other states, NH is using what’s called a “Student Growth Percentile” model that is similar to value-added models.  It takes into account students’ starting points and the progress they have made during the year.

The main difference between the approach being used in NH and the methods advocate by Rhee is that NH recognizes how difficult it is to attribute the change in student learning to any single teacher.  Many adults influence a student’s learning, even in apparently self-contained classrooms like elementary classes.  So the NH Student Growth Percentile analyses are shared among all the teachers of a particular group of students.  For example, a particular school may choose to share the Student Growth Percentile results of the fourth grade students with all of the fourth grade teachers and specialists (reading specialists, for instance).  This avoids many of the issues raised by other value-added models.

Even at that, only 30% of the teachers scores evaluations can be calculated using state test scores.  The remaining 70% of teachers are in subjects for which there are no standardized tests.  The NH Task Force has recommended setting Student Learning Objectives in these subjects.  Teachers would set Student Learning Objectives for, say, a school year, and check the progress students are making, and then evaluate whether their students achieve those goals.  We’ll discuss Student Learning Objectives more in future posts.

Evaluating teachers is a complex and, now, politically fraught topic, but we seem to be taking a thoughtful approach here in New Hampshire.