D.C. Public Schools is planning to reduce the power that students' standardized test scores have over teachers' performance ratings and job security, The Washington Examiner has learned.

In a major change to its controversial Impact teacher evaluation tool, the school system also will reduce the number of classroom observations teachers must undergo, and add a new performance category to recognize "developing" teachers, internal sources said.

Melissa Salmanowitz, a spokeswoman for Chancellor Kaya Henderson, confirmed that D.C. Public Schools plans to announce changes to Impact Friday morning.

History of changes
Friday's announcement won't be the first major change to Impact since its introduction in 2009.
• Chancellor Kaya Henderson tweaked Impact in its second year into "Impact 2.0," which she described as more rigorous than the original evaluation. The change made the evaluation framework more flexible, but streamlined. It also added a "commitment to school community" rubric and cut some elements frustrating to teachers.
• A law passed by the D.C. Council this spring requires the city to establish a pilot program to draw top teachers to low-performing schools with $10,000 bonuses, housing assistance, income tax credits and other incentives. While this does not directly change Impact, it certainly adds to the picture.
• This school year, D.C. Public Schools experimented with fewer observations for teachers. The Washington Examiner reported in September that teachers who had been top-rated for the past two years, and who earned high scores on their first two evaluations in the fall, would have the option of waiving the remaining three observations. Nearly every teacher did.

Currently, teachers whose classrooms are tested on the D.C. Comprehensive Assessment System receive performance evaluations based 50 percent on their students' improvement on the exams. Sources tell The Examiner that only 35 percent of the ratings will now come from the test scores.

Since 2009, about 400 D.C. teachers have been fired for poor evaluations, including 98 this week. The evaluations' emphasis on testing has been the most controversial aspect of Impact, developed by Henderson when she was the deputy to former Chancellor Michelle Rhee. A handful of teachers have been fired for correcting their students' bubble-sheet answer to boost their evaluations.

The Examiner first reported in April that DCPS was considering scaling back the weight given to standardized test scores on teacher evaluations. In May, the Office of the State Superintendent for Education urged schools to fill in the gap with SAT scores, end-of-course exams and other performance measures.

D.C. Public Schools is also planning to reduce the number of classroom observations, which some teachers find nerve-wracking, from the current five. Observations count toward 35 percent of the evaluations of teachers whose students take the state exams, and 75 percent of the evaluations in other general-education classrooms. Sources say one of the observations will be "formative" and not count toward the Impact score.

Washington Teachers' Union President Nathan Saunders said he supports the test-score and observation tweaks, but is concerned about the new classification being added to the evaluations. Currently, teachers receive ratings of "ineffective," "minimally effective," "effective" and "highly effective."

Saunders said he has shared concerns with Henderson about the "developing" category being added, which would recognize new teachers who are still adjusting. The new label would increase the evaluation score needed to receive an "effective" rating, putting the positive label further out of touch. Teachers who are "developing" for three years could be fired.

"I dont believe you should punish [teachers still developing] with the potential of unemployment," Saunders said. "When I learned of it, I honestly tried to talk [Henderson] out of it."

Salmanowitz declined to comment on the details of the changes because D.C. Public Schools had promised an exclusive story to The Washington Post.