How not to look for Grade-A teachers, Peter Smyth

How not to look for Grade-A teachers, Peter Smyth

Charleston Post and Courier
March 13, 2013

[click to see larger]

How not to search for grade-A teachers

The editors of the Post and Courier on March 7 called for a "Search for grade-A teachers." They supported the South Carolina system for teacher evaluation, which would use standardized test scores as a measure of student growth as 30% of a teacher's evaluation. This is called the Value Added Method, or VAM. This is a bad idea for many reasons, but it is flawed as a strategy to get at what the editors seem to want.

While observations would comprise 70% of the evaluation, the VAM score in many cases would become the tie breaker. A very basic problem with observations is the time it takes to observe and do he followup work. Principals just don't have the time to focus on all teachers. So the 30% VAM is the part that really counts.

The flaws in the VAM have consequences. In New York, a very similar system had a margin of error of close to 50% on a 100% scale. That means we can be confident that a teacher could be anywhere from highly effective to ineffective.

So for a teacher, three things could happen. The rating gets the true effectiveness correct. With the high margin of error, flipping a coin would do as well and save time.

The rating could be a false positive: an ineffective teacher incorrectly gets a high rating. That means the teacher is even further entrenched. Once that teacher has a high rating, right or wrong, he is locked in, and with that margin of error, this will happen.

The rating could be a false negative: a truly effective reacher could be rated as ineffective. The real consequence here is that more good teachers, the ones we are searching for, will leave the profession. That's not just bad for the teacher, it's bad for students.

Thus the editors, Mick Zais, Nancy McGinley, Arne Duncan, and others support a system for teacher evaluation that will likely entrench more ineffective teachers while reducing the number of effective teachers. That's worse than regression to the mean. They back a system with no supporting evidence, but only a nice story line and the hope it might work. Sounds like a plan.

The whole issue is more deeply flawed because it defines what we want for our children and what we want teachers to focus on, as increasing scores on narrow standardized tests. One would think we might look at other countries with successful education systems, or closer to home, certain private schools.