The president of the American Federation of Teachers has proposed a set of policies for changing the way that teachers are paid. Some of it sounds good, some of it laughable, and some of it suspicious, but overall this sentiment is very encouraging:
“Our system of evaluating teachers has never been adequate,” she said, adding that for too long it “has failed to achieve what must be our goal: continuously improving” teaching.
The proposals, from the AFT website, include setting standards for what teachers should “know and be able to do”. To me, this sounds like more certification and education requirements, which the evidence shows does not impact teacher ability.
In addition, they propose that teacher assessments should be based on a variety of metrics including in-class observations, self-evaluations, portfolio reviews, and assessments of lesson plans, students’ work, and “other projects”. My first reaction is that self evaluations are a laughable means by which to judge teachers, but that more measurements are probably a good thing.
They also want test scores to be used only to compare scores of students from the beginning of the year that they are with the teacher to the end of the year. Allowing the teachers to administer their own baseline tests provides them with terrible incentives. I’m not sure why this measurement would be better than comparing to students’ previous year’s scores except that it allows teachers to influence their own baseline.
Another reform includes hiring Ken Feinberg, Obama’s Compensation Czar, to oversee develoment of a new process for removing bad teachers.
Overall I find it encouraging that the reform proposals of one of the nations largest teachers unions recognize that it needs to be easier to fire bad teachers, and that some form of standardized testing is a good thing.

2 comments
Comments feed for this article
Wednesday ~ January 13th, 2010 at 11:32 am
dWj
This is absolutely a step forward coming from a teachers’ union. (The administration is already far ahead of this, but I’m grading on a curve.)
My problem with basing measurements on “in-class observations, self-evaluations, portfolio reviews, and assessments of lesson plans” — input-based measurement — is that I’m skeptical that the people who would be doing these assessments are omniscient about what constitutes a proper in-class dynamic or presentation, or what the optimal lesson plan for a particular class of students is. Leaning mostly on output-based measurement both improves the extent to which we’re measuring what we want to measure and creates incentives for the teachers to experiment and try to improve on previously established practices.
In http://www.theatlantic.com/doc/print/201001/good-teaching it is noted that the teachers that get the best results, at least among those these people investigated, were routinely tearing things up and rebuilding them. We want to encourage that.
Monday ~ January 18th, 2010 at 9:00 am
Links and Quotes « Modeled Behavior
[...] Seyward Darby at TNR is less enthusiastic than I was about the American Federation of Teachers new reform proposals. He’s probably right: In many [...]