The third and final set of results from the MET study was released last week and it packed a solid punch. Specifically, the research helps answer a crucial question: What are the key factors in designing a teacher evaluation system?
Design demands trade-offs. Evaluation systems must balance a number of competing aims: identifying teachers who get their students to do well on state tests, reliability, efficiency, and providing opportunities for feedback. A state or district that wants to maximize how well its system predicts which teachers will get the biggest gains on state tests, according to MET, should set the weight given to test scores at 50% (or even higher). But a system that gives at least a third of the weight to observations of teaching provides greater reliability -- in other words, more stability in the rating assigned to teachers from year to year. Such a system would be more likely to earn the trust of teachers on whom, ultimately, the success of the schools depends.
The MET study's findings about efficiency -- a crucial consideration when states and districts grapple with implementation -- are of particular interest. MET shows that observations adding up to 180 minutes are only marginally more reliable than one 45-minute and 3 15-minute observations performed by a school administrator and other observers.
After a certain point, increasing the amount of time to observation yields significantly diminished returns. It's heartening that districts and teachers can feel confident that a manageable amount of time devoted to each teacher's evaluation can still provide trustworthy results.
So who's ahead of the curve? Come back to this space tomorrow and we'll have information on which districts are using peer and third party evaluators as well as student feedback in January's edition of our Tr3 Trends.
--Sudipti Kumar