Using kids' test scores, DOE conducting secretive experiments on teachers

Posted by Philissa at 9:52 AM

Who knew I was already right when I hypothesized two weeks ago that the DOE was hoping to change the way teachers are evaluated? Well, besides Eduwonkette, who left a comment telling me so, and at least 140 principals whose teachers are already being judged according to their students' test scores in an initiative so top secret that even the teachers don't know about it? Very few people, it appears, according to the New York Times.

In the already-underway experiment, which the Times was the first to report, the test score gains of students at 140 schools will be used to judge their teachers' success. The DOE is setting "predicted gains" for teachers based on their students' skills, experiences, and backgrounds — and then crunching the numbers to see if the teachers meet those goals. The DOE told the Times, which broke the story, that it doesn't plan to use the results to make hiring or firing decisions about individual teachers. But Chris Cerf, who apparently has been deputized to talk up the program, said the results could be one factor used in those decisions, and that ultimately making the results public (a la the progress reports) would reward good teachers and put pressure on bad ones. Certainly, the DOE must be interested in providing more ammunition for the teacher firing squads assembled earlier this year.

Naturally, the UFT's Randi Weingarten, who has backed down in her opposition to other controversial plans, including the Teacher Performance Unit, sounds angry about this one, telling the Times that she and the city disagree on whether results from this pilot or its expansion could be used under the teachers' contract to make hiring or firing decisions. (On the other hand, the Times says the UFT has known about the experiment for four months, but we haven't heard any complaints until now.)

The initiative also appears to undercut the little agency afforded teachers in determining how performance pay is distributed this year. I'm pretty sure that we don't know how many of the schools included in the performance pay pilot elected to distribute their earnings across the whole faculty rather than to individual teachers, but I think it's safe to guess that's what happened in most schools. Now the DOE is doing the divisive, problematic work its teachers declined to do.

The Times predicts a battle this summer between the DOE and the UFT over the experiment results. Let's hope Randi Weingarten (or, potentially, her successor) is up for the fight. The DOE is abusing test score data, which aren't meant for this kind of crunching, and keeping teachers in the dark about how they're being evaluated. Regardless of the quality of the research (though even that is questionable — Eduwonkette wonders whether the experiment is ethical given that many of the research subjects don't know they are part of an experiment at all), the way the DOE has gone about this one is just not right.