How many New York City public schoolteachers are so incompetent that they should be fired? That’s the 250-million-dollar question that must be addressed by both sides wrangling over what kind of teacher-evaluation system the city is going to build.

New York City Mayor Michael Bloomberg (photo by David Shankbone)

For months now, despite a state mandate to build such a system, Mayor Michael Bloomberg and the city’s Department of Education have been locked in a stalemate with the United Federation of Teachers over the terms of a teacher-evaluation process that, by law, must be agreed to via local collective bargaining.

The parties have already missed a Jan. 17 deadline set by the governor, sacrificing a 4 percent increase in state aid for education to the city. But the governor and other state officials should have known that punishing the city and its children by withholding this aid—and future funds as well—would be both bad public policy and an ineffective strategy to force an agreement.

This dispute is about principles that each side believes to be far more important than the money at stake, and at the heart of the disagreement is just how many teachers we’re talking about calling incompetent—and therefore unsuited to educating our kids.

The mayor’s stance is clear: there are good teachers and bad teachers in New York City, and a teacher-evaluation system must be able to identify and fire consistent low-performers. Their replacements are likely to be better teachers, and children’s achievement will increase as a result. If these new teachers have lower salaries than the ones they are replacing, so much the better.

Published reports suggest that Mayor Bloomberg quashed a tentative agreement between the Department of Education and the union due to a “sunset” clause providing that the agreement would be in force for two years and then be renegotiated. Indignant, the mayor said that any agreement that would end before the completion of a process to remove a teacher rated “ineffective” in two consecutive years was a sham. If the agreement would expire before any teacher had been fired for cause, he argued, it was worthless.

The union’s stance is equally clear: a teacher-evaluation system must be fair to teachers, providing an accurate picture of their performance and an opportunity to improve if their performance is not up to snuff.

The issue of accuracy is a serious bugaboo: estimates of teachers’ contributions to student learning, whether in the form of value-added measures or growth percentile scores, are imprecise, and, if they fail to take account of factors beyond a teacher’s control, unfair. Moreover, many teachers fear that an unscrupulous or incompetent principal will rate a teacher unfairly when observing him or her in the classroom. The union thus seeks to build procedural safeguards into the evaluation system to minimize the risk that a teacher will be unjustly identified as ineffective, and subsequently terminated.

Lost in this clash of principles—efficiency vs. fairness—is the question of just how many teachers ought to be judged ineffective and fired.

Is it two percent? Five percent? Twenty percent?

If you believe our schools are failing kids across the board, you’d likely set the number high, citing the very low number currently fired as utterly unacceptable. But if you’re a parent, typically satisfied with your child’s own school and teacher, you’re likely to set the number low. If you’re Mayor Bloomberg, you might covet the flexibility and cost savings that would come with dismissing a large number of highly paid teachers and replacing them with malleable and cheaper novices. And if you’re the teachers union, your legitimacy is in the hands of your members, none of whom wants to be fired, especially for reasons they deem unfair.

The truth is that, no matter how much we try to craft teacher-evaluation systems that are fair and impartial, the question of how many teachers should be rated “ineffective” and dismissed is still a value judgment. And that fact reveals just how arbitrary the new world of rating teachers can be.

A little over a year ago, Bloomberg told a group of students at M.I.T. that given the opportunity, he’d fire half of New York City’s teachers and double the compensation and class sizes of the remaining “good” ones. That’s a lot more than the 18 percent that an evaluation system piloted at 20 New York City schools identified as “ineffective.”

But 18 percent is a big number, too. In Washington, D.C., which has pioneered an evaluation system similar to the one New York City might adopt, two percent of teachers are rated “ineffective,” and an additional 14 percent are rated “minimally effective”—but two consecutive ratings of “minimally effective” can result in termination.

In New Haven, Conn., which also uses a system similar to what’s being developed in New York, 10 percent of teachers received ratings of “needs improvement” or “developing,” the two lowest categories in a five-category evaluation system. Hillsborough, Fla., the nation’s eighth-largest school system, initially proposed that at least five percent of tenured teachers would be dismissed each year for poor performance under its new teacher-evaluation system. But with experience, the district has determined that only about 1.5 percent of teachers are unsatisfactory, and another two or three percent receive a rating of “needs improvement.”

Are we really to believe that the quality of teachers varies so dramatically across districts and over time? It’s far more likely that these figures reflect local values and priorities. And this is the heart of the New York City dispute. Mayor Bloomberg prizes efficiency and believes that the discretion to fire a large number of teachers is essential. The UFT champions fairness and due process, asserting that teachers who’ve been awarded tenure by the city have demonstrated their effectiveness, and that very, very few have received unsatisfactory evaluations from their principals. The values conflict boils down to a big number versus a small number.

As Bloomberg’s tenure as mayor winds down, he’ll seek to cement his legacy, with education as a signature issue. An evaluation deal that is too easily undone will tarnish that legacy. Conversely, the UFT, aware of the mayor’s disdain for the teachers of New York City, is content to wait him out, betting that a new mayor will be more favorably inclined toward teacher evaluation, and perhaps working conditions and compensation as well.

Neither side is likely to agree to an evaluation system that gives the other party too much control over who gets fired. Perhaps the current standoff wouldn’t be intractable if the two sides could at least agree on some bounds for the number of teachers who will be judged “ineffective” and subject to termination. But what’s the magic number?