City teachers' performance ratings public

New York City took the unprecedented step Friday of releasing the names and job-performance rankings of nearly 18,000 public school teachers -- but with a caveat that many ratings are statistically flawed and potentially misleading.

The acknowledgment of errors confirmed the apprehensions of educators on Long Island, who have strongly protested state plans to calculate ratings for thousands of teachers in Nassau and Suffolk districts by June. State Education Department officials said Friday that teachers will have a chance to verify information in evaluations prior to release.

A wrinkle in the city ratings was that they were based on student improvement on standardized state English and math tests over one or more years. That required detailed analysis of the students involved -- for example, their poverty levels and rates of absence from school -- that analytical experts say are frequently subject to mistakes.

As one example of errors in the city teacher ratings, school officials said teachers who had checked class rolls used in evaluating performance identified, on average, one student each who they hadn't actually taught. Underlining the potential for mistakes, city officials added that only 37 percent of teachers had verified the class rolls included in their evaluations.

At a 90-minute news briefing Friday in Manhattan, city school officials cautioned that the teacher rankings being released were 2 years old and, essentially, only estimates of teacher performance. For example, they said, the average statistical "margin of error" for teacher ratings based on students' math scores was 35 points.

Simply put, that means teachers who ranked in the middle, or 50th percentile, among colleagues actually might merit ratings ranging between the 32nd and 67th percentiles. A separate Newsday analysis found 86 teachers with error ranges of 80 points or more in 2009-10, the latest year calculated.

"No principal would ever make a decision based on this score alone," said Shael Polakow-Suransky, the city's chief academic officer, referring to the rankings.

City school officials emphasized that they had used other approaches, including classroom observations, to develop a more rounded picture of teacher performance. Of 133 teachers flagged due to low statistical rankings last year, 36 percent ultimately were granted job tenure, those officials said.

On a morning radio show, city Mayor Michael Bloomberg acknowledged the limitations of the ratings, but defended the use of test scores to help decide teachers' job status.

"This school system is run for the kids, and you can't leave the kids in front of a teacher who isn't doing the job for any length of time," he said.

The ratings drew sharp rebukes from teacher union leaders and others.

"This was a complete calamity," said Michael Mulgrew, president of the 200,000-member United Federation of Teachers.

"Anyone that tries to use this negative information to harm a teacher, we're going to stand by that teacher."

"The public shaming of teachers that accompanies this release . . . will only serve to undermine their usefulness in the eyes of teachers and the general public," said Sean Corcoran, an associate professor of educational economics at New York University.

Randi Weingarten, head of the American Federation of Teachers, said the ratings' release will negatively affect the way teachers regard the collection and use of such information.

"Instead of teachers embracing the use of such data, it will be viewed as a weapon," she said. "It's also one more step in encouraging this testing fixation as opposed to turning away from it."

The ratings rank teachers on a statistical curve -- meaning that certain numbers automatically fall in the bottom 5 percent, the top 5 percent and so forth.

Starting in June, teachers in the city, on the Island and elsewhere will be rated under a different state system that sets no numerical quotas, but separates teachers into four categories ranging from "highly effective" to "ineffective." Those deemed ineffective two years running face possible job loss.

State education officials said they're still weighing the question of whether to release teachers' names along with ratings. On the Island, many educators regard release of names as inevitable, due to public pressures for government accountability.

"Given what we've seen so far from the politicians in charge, I have no doubt that, if they think they can get public traction through release of names, that they'll do it even if they know the information is inaccurate," said Arnold Dodge, chairman of the Department of Educational Leadership and Administration at LIU Post in Brookville.

The United Federation of Teachers fought in court to block release of the ratings, contending that testing data were seriously flawed. But the state Court of Appeals last week upheld a lower-court ruling that teachers' names and test results could be made public.

Several news media outlets had sought the records' release under the Freedom of Information Law. Newsday was not a party to that lawsuit.