In New York City — "District D" — there was virtually no difference in turnover rates based on teacher performance.

Getting rid of weak teachers doesn’t always require massive policy changes. Sometimes all it takes is a nudge, a new study on teacher turnover suggests.

When New York City principals told low-rated teachers that they were deficient, the teachers were three times more likely to leave the school, according to the study, released today by TNTP, a group that advocates for aggressive changes to hiring and firing practices in public schools.

Convincing the best teachers to stay is just as easy as counseling the weak ones out, the study suggests. Top-rated teachers said they were more likely to stay if their principals gave them more constructive feedback and more public recognition for their efforts, but two-thirds of them reported that their principals did not even encourage them to return to their school.

The study is a follow-up to TNTP’s 2009 influential “Widget Effect” report, which urged school districts to revamp teacher evaluations. In the new report, the group focuses on how districts can hold on to teachers determined to be the best. Districts don’t make a special effort to keep those teachers, termed “Irreplaceables” in the report, and when they leave, schools are highly unlikely to hire teachers who are anywhere near as strong, the report concludes.

Some of the report’s findings represent low-cost, easy-to-implement alternatives to some of the other policies TNTP has pushed, including firing teachers who don’t have permanent positions and doing away with seniority-based layoffs.

TNTP researchers examined test score data for 90,000 teachers in four large urban districts and one charter school network. The study then linked performance data to surveys that teachers took to find out what would keep top-scoring teachers in their schools and what could cause low-scoring teachers to leave.

One data set the report used was New York City’s short-lived “value-added” teacher ratings, issued in 2008-2010 to some elementary and middle school teachers. When the ratings went public in February, as work on the report was underway, city officials and their critics alike warned that the volatile scores were not accurate measures of teacher quality.

The study recommends that principals should be evaluated based on turnover among their top-performing teachers, a proposal that is theoretical right now in districts such as New York City that do not have sophisticated evaluation systems in place. The study also recommends that the best teachers should earn more than $100,000 if they are still working in a high-needs urban school after six years.

New York City has never proposed bumping teacher salaries up so fast. But Mayor Bloomberg did propose a bonus plan in January that would give top-rated teachers a $20,000 raise. The city teachers union, which staunchly opposes merit pay for teachers, quickly passed a resolution opposing the proposal.

Today, Bloomberg and Chancellor Dennis Walcott blamed the teachers union for blocking their retention proposals.

“Unfortunately, the United Federation of Teachers continues to stand in the way of these incentives,” they said. “It’s time for the union to work with us so that our irreplaceable teachers can take full advantage of these and other opportunities.”

UFT President Michael Mulgrew criticized Bloomberg for ignoring some of the report’s other recommendations, such as improving working conditions in high-needs schools. ”It’s a shame that the mayor, who thinks merit pay is the solution to every problem, has chosen to ignore one of this report’s central findings — that ‘poor school cultures and working conditions drive away great teachers,’” he said in a statement.

Mulgrew also criticized the report’s reliance on test scores to identify the best and worst teachers.

TNTP researchers acknowledged that they would have preferred to evaluate teachers on more than just test scores, but that was all they had. TNTP President Tim Daly said he was confident using the ratings because a previous study conducted by the Gates Foundation found a strong link between classroom observations and students’ test score gains.

TNTP was founded by former Washington, D.C., Chancellor Michelle Rhee, a controversial national figure whose efforts to change how teachers are hired and fired has antagonized teachers unions. TNTP, which handles the recruitment and training of the New York City Teaching Fellows, also has a robust research arm that regularly studies the teacher job market across the country.

Overall, the study found that districts typically retain strong and weak teachers at the same rates: Six percentage points separated the average turnover rates among the highest- and lowest rated teachers in the districts studied. In New York City, for instance, schools retained 89 percent of their top performers and 88 percent of their bottom performers during the 2009-2010 school year.

“The real teacher retention crisis is not simply the failure to retain enough teachers; it is the failure to retain the right teachers,” the TNTP researchers wrote.

“It’s tragic we haven’t been doing easy things to ID & retain teaching talent–& inexcusable if we don’t change our ways,” U.S. Education Secretary Arne Duncan tweeted today about the study’s findings. Duncan participated in a panel discussion in Washington, D.C., about the study.

Daly said the study’s biggest revelation was that targeted teacher retention could improve with even the slightest feedback from principals.

Daly said that in the early stages of the study, his research team informally polled union officials and policy-makers about what they thought could improve teacher retention. One hypothesis the researchers tossed out was better principal support or feedback.

“I can’t tell you how many times we brought that up and they almost always questioned whether that even makes a difference,” he said.

But Daly said TNTP’s findings should force policy makers to think differently about how they approach the issue.

“The conventional wisdom about what works is frequently wrong,” Daly said.

TNTP Irreplaceables Study — July 30, 2012