“Teachers have to be parents, priests, lawyers, clothes washers, babysitters and a bunch of other things” if they work with low-income children, said Nathan Saunders, president of the Washington Teachers Union. “Impact takes none of those roles into account, so it can penalize you just for teaching in a high-needs school.”

Jason Kamras, the architect of the system, said “it’s too early to answer” whether Impact makes it easier for teachers in well-off neighborhoods to do well, but pointed out that Washington’s compensation system offers bigger bonuses ($25,000 versus $12,500) and salary enhancements in high-poverty schools.

“We take very seriously the distribution of high-quality teachers across the system,” he said.

The evaluation system leans heavily on student test scores to judge about 500 math and reading teachers in grades four to eight. Ratings for the rest of the city’s 3,600 teachers are determined mostly by five classroom observations annually, three by their principal and two by so-called master educators, most recruited from outside Washington.

For classroom observations, nine criteria — “explain content clearly,” “maximize instructional time” and “check for student understanding,” for example — are used to rate the lesson as highly effective, effective, minimally effective or ineffective.

These five observations combine to form 75 percent of these teachers’ overall ratings; the rest is based on achievement data and the teachers’ commitment to their school communities. Ineffective teachers face dismissal. Minimally effective ones get a year to improve.

Impact costs the city $7 million a year, including pay for 41 master educators, who earn about $90,000 a year and conduct about 170 observations each. The program also asks more of principals. Carolyne Albert-Garvey, the principal of Maury Elementary School on Capitol Hill, has 22 teachers — she must conduct 66 observations, about one every three school days.

“I’ve really gotten to know my staff, and I’m giving teachers more specific feedback,” Ms. Albert-Garvey said. “It’s empowered me to have the difficult conversations, and that gives everyone the opportunity to improve.”

Advertisement Continue reading the main story

Several teachers, however, said they considered their ratings unfair.

A veteran teacher who said he did not want to criticize the school system openly, said that a month after he inherited a chaotic world history class from a long-term substitute, the visiting evaluator cut him no slack for taking on the assignment and penalized him because a student was texting during the lesson.

Another teacher who expects to lose her job next month because of low ratings said at a public hearing that evaluators picked apart her seventh-grade geography lessons, making criticisms she considered trivial. During the most recent observation, her evaluator subtracted points because she had failed to notice a girl eating during class, the teacher said.

“I’m 25 years in the system, and before, I always got outstanding ratings,” she said. “How can you go overnight from outstanding to minimally effective?”

A report issued by the Aspen Institute in March said one of Impact’s accomplishments was to align teacher performance with student performance, noting that previously 95 percent of Washington’s teachers were highly rated but fewer than half of its students were demonstrating proficiency on tests. Still, the report quoted teachers who complained of cold-eyed evaluators more interested in identifying losers than in developing winners.

“After my first conversation with my master educator, I felt it was going to be worthwhile — she offered me some good resources,” the report quoted one teacher. “My second master educator was kind of a robot, not generous in offering assistance, a much tougher grader.”

This month, Mary Gloster, who taught science in three states before she was recruited to Impact in 2009, was at Ballou High, one of the city’s lowest-performing schools, to share the results of some classroom visits.

She met with Mahmood Dorosti, a physics teacher who won a $5,000 award this spring. “Don’t even think about it — you’re highly effective,” she told him.

Next was Ms. Strzelecki, 23, who came to Ballou through Teach for America . The two sat at adjoining desks, with Ms. Strzelecki looking a bit like a doe in the headlights.

Advertisement Continue reading the main story

But Ms. Gloster, who had watched her teach a ninth-grade biology lesson the week before, offered compliments, along with suggestions about how Ms. Strzelecki might provide differentiated teaching for advanced and struggling students.

“You did a really good job, kiddo,” the evaluator ruled, grading her as effective, the equivalent of a B (the same rating she got on previous observations).

“What I liked about Mary was that I felt she was on my side,” Ms. Strzelecki said later. “Some teachers feel the master educators are out to get them.”

That is a common perception, said Mark Simon, an education analyst for the Economic Policy Institute , which receives teachers’ union financing. Ms. Rhee developed the system, he noted, during tough contract negotiations and did not consult with the teachers’ union in its design.

“That was a missed opportunity,” Mr. Simon said, “and it’s created a lot of resentment.”