Teachers can expect unannounced observations to factor into their annual ratings under the terms of the evaluations agreement that Gov. Andrew Cuomo announced today.

The unannounced observations are one of several ways that the State Education Department and state teachers union, NYSUT, agreed to flesh out the state’s 2010 evaluation law, seen as so open-ended as to stymie implementation.

The agreement, which Cuomo is set to turn into law through the state budget amendment process, resolves some major points of contention while continuing to leave many elements of districts’ evaluation system subject to local collective bargaining. Districts and their unions have until the end of 2012 to turn the framework into a local evaluation system, or risk losing state aid.

The framework hews to the broad contours of the 2010 teacher evaluation law: 20 percent of ratings will be based on a calculation of student growth based on state test scores; 20 percent will be based on other assessments that are decided locally; and 60 percent will come from subjective measures such as observations, also decided upon locally. Teachers will still receive a score between 0 and 100 and a rating ranging from “ineffective” to “highly effective.” But there are new constraints.

In a major win for the state, teachers whose students show no academic growth will get an “ineffective” rating, even if the rest of their evaluation is strong. The evaluation law had not provided for such a circumstance.

The 20 percent of ratings that are supposed to come from local assessments had been the most obvious point of contention between SED and NYSUT. Last year, NYSUT sued the state after Cuomo pressed the Board of Regents to double the weight of student test scores in teachers’ evaluations. Today’s agreement does allow districts to use state test scores for the local assessments — as long as their calculation of student growth isn’t based on the same formula as the state’s.

For example, districts could crunch the state’s numbers to show the growth of students in high-needs groups, or they could calculate an entire school’s test score climb to complement that of a single teacher’s students. Districts can also elect to use assessments produced by third-party vendors or create their own assessments, something the city had been working on when teacher evaluation negotiations fell apart at the end of December. Today’s agreement gives the state the right to vet the rigor of local assessments.

Richard Iannuzzi, NYSUT’s president, said the agreement announced today adequately constrained the role of test scores.

“While there is a place for standardized testing in measuring teacher effectiveness, tests must be used appropriately,” he said in a statement.

The agreement also clarifies the role of observations in a teacher’s rating. Previously, the evaluation law didn’t actually say that teachers would have to be observed as part of the evaluation process. The new framework guarantees that at least 31 percent of the ratings is based on at least two observations by principals. One of those observations must be unannounced, something that teachers unions have long opposed.

The agreement also bolsters the role of the state education commissioner. Now, the commissioner will have the right to reject local evaluation systems on grounds other than whether they simply comply with the state law. Systems that appear to crafted outside the spirit of the law — to make evidence of student growth a crucial component of teacher ratings — are also subject to rejection.

And the commissioner is also getting the right to evaluate teachers if he or she suspects their districts are not doing a good job.   The provision is meant as a safeguard against the current reality, in which virtually all teachers across the state receive “satisfactory” ratings each year even as student achievement results suggest that schools have a great deal of room to improve. State officials said they had no goals for the number of teachers who receive low ratings under the new system and did not anticipate stepping in to issue new evaluations for individual teachers in most cases.

Today’s agreement represents a substantial narrowing of the world of possibilities under the state’s evaluation law. But much of what will actually make up teachers’ ratings remains up to negotiation between local districts and their unions. Both the local assessments and the subjective measures other than observations must be bargained. Negotiations will direct what observation model principals when watching teachers in their classrooms. And, of course, districts and their unions will have to agree on the appeals process for teachers who get low ratings. That element had been the sticking point in New York City, and a separate deal on that front announced today won’t go into effect until all of the other pieces fall into place.

For the hundreds of school districts across the state that had actually come to an agreement on new teacher evaluations, the framework is sending them back to the negotiating table. Their deals will remain in place for the remainder of this school year, but next year they will have to set new evaluation systems that match the state’s updated framework.