An experienced teacher asked me to cover her class for a few minutes, she came back looking pale, I asked her if she felt okay, she told me she went into the bathroom to throw up, she was being observed the next period.
A principal called to cancel a meeting, “I have to do teacher observations for the next few days.”
I asked, “Are they useful?”
“Are you kidding, a total waste of time, a mechanical exercise.”
Teachers hate to be observed and school leaders feel it’s a waste of time.
In a handful of schools across the city the observation process is actually linked to instructional practice. Teachers, utilizing a checklist, observe colleagues, meet and discuss the lesson.
One principal encourages online responses to observations and creates a dialogue about kids and practice.
In one department in a large high school experienced teachers can opt for “collegial” observations, experienced teachers observing each other in lieu of traditional observations.
How can we improve the system so that teachers and school leaders see it as a process to improve instruction as well as to evaluate practice?
Section 3012-b of the Education Law (as added by Chapter 57 of the Laws of 2007) resulted in new standards and procedures for making tenure determinations for teachers in the instructional services employed in school districts and BOCES. The new standards for making tenure decisions allow for “peer review by other teachers, as far as practicable.”
When UFT President Mulgrew asked teachers at a monthly Delegate Assembly meeting how they felt about peer review, he received a very lukewarm response.
The vacuum has been filled by Arne Duncan, use student achievement data to assess/evaluate teachers.
Ironically as Duncan sprints around the country lauding assessment based teacher evaluation system his own Institute for Educational Sciences warns,
Simulation results suggest that performance estimates are likely to be noisy using the amount of data that are typically used in practice—1 to 3 years. Type I and II error rates are likely to be about 25 percent based on three years of data and 35 percent based on one year of data.
Aaron Pallas in his Hechinger Report blog agrees that there basically is no meaningful teacher evaluation at the local level but warns about overemphasizing student achievement data.
That upwards of 95 percent of all teachers receive fully satisfactory ratings, with little additional information in the evaluation on how to improve performance, is indefensible ….
I’d like to articulate a principle: the weight accorded to any one element of a teacher evaluation system should be proportional to the uncertainty about the inference which is drawn from that element. It’s a cardinal rule in social science research: acknowledge the uncertainty about the fit between the data and the claims a researcher seeks to make based on those data. When there is more confidence that an element is a “good” measure – that is, a reliable and accurate representation of what it’s supposed to measure – give that element more weight. When there is less confidence that an element can support an inference, give it less weight.
Pallas goes on to explain why the LA Times data simply doesn’t pass the “valid and reliable” test, see the reasons here.
So, if teachers and supervisors abjure the traditional system and the current VAM data systems are deeply flawed where do we go?
* Teachers and school districts, in a collaborative fashion, must design their own teacher evaluation systems, and, with the help of a grant from the AFT a number of local unions and school districts are designing local options,
Local union and school district teams from Albany, Hempstead, Marlboro, North Syracuse and Plattsburgh are developing the critical components of a new evaluation option, with guidance from national experts.
* Teachers, and their unions, must accept more responsibility through peer review, teachers playing a role in the assessment/evaluation of colleagues. In the past some unions has entered the peer review arena, however, without strong endorsement by their memberships the vacuum will be filled by data-driven evaluation systems.
* We must choose principals who are strong classroom teachers and have demonstrated leadership experience. The trend to select non-educators or new teachers is foolish. The Leadership Academy in NYC, and the New Leaders for New Schools graduates are resented and not respected by classroom teachers.
I was talking with a retired school supervisor and asked him how many teachers in his school were truly terrible. Over his last five years out of a couple of hundred teachers he remembered three, one of whom was discharged.
He said, “We had more than our share of stars, mostly hard-working conscientious teachers and a few duds.”
He is absolutely correct. The vast majority of teachers (within two standards deviations of the mean, for the data mavens) are incredibly hardworking practitioners, their “effectiveness” varies from year to year. We should investigate the qualities of the “stars,” and offer assistance to, and move to discharge the duds, maybe the 2% at either end of the spectrum.
The unintended consequences of the fixation over teacher quality/teacher dismissal is to alienate teachers on the political side , discourage risk-taking in the classroom and encourage teaching to the test.
Bad outcomes for all.