When Terry Grier was hired to run the San Diego Unified School District in January 2008, he hoped to bring with him a revolutionary tool that had never been tried in a large California school system.
Its name -- "value-added" -- sounded innocuous enough. But this novel number-crunching approach threatened to upend many traditional notions of what worked and what didn't in the nation's classrooms.
Rather than using tests to take a snapshot of overall student achievement, it used scores to track each pupil's academic progress from year to year. What made it incendiary, however, was its potential to single out the best and the worst teachers in a nation that currently gives virtually all of them a passing grade.
In previous jobs in the South, Grier had used the method as a basis for removing underperforming principals, denying ineffective teachers tenure and rewarding the best educators with additional pay.
In California, where powerful teachers unions have been especially protective of tenure and resistant to merit pay, Grier had a more modest goal: to find out if students in the district's poorest schools had equal access to effective instructors.
Still, it proved radioactive to San Diego's teachers union. Like many unions across the country, it saw the approach as a flawed instrument, a Trojan horse for introducing merit pay and a threat to hard-won employment protections.
After nearly two years of grinding battles with the union and school board on this and other issues, Grier recently left for Houston, where the district uses value-added results as a basis for teacher bonuses.
The opposition in San Diego, Grier said mildly, was "more entrenched than I thought it would be."
His fight there offers a preview of a debate that is about to engulf the nation's schools.
The Obama administration has made value-added a pillar of its school-reform efforts, including the $4.35-billion federal grant program known as Race to the Top, which requires states to link student scores to teachers.
The administration's endorsement thrust to the fore a 20-year-old idea that has long been confined to academic circles and a handful of states and school districts, including Chicago's, where it was championed by now-Secretary of Education Arne Duncan.
Gov. Arnold Schwarzenegger this month signed into law two measures that put California on the path to measuring student growth with the value-added method.
But most districts thus far have just begun to talk about using the method in teacher evaluations.
Value-added analysis promises to address one of public education's central conundrums: On one hand, research shows that effective teachers are the single most important factor in improving student performance. On the other, most states use subjective evaluation systems -- based on occasional classroom visits by administrators -- that give nearly all teachers a satisfactory rating.
"Zero percent [of teacher evaluation] is based on student achievement," Duncan said in a recent interview. "That's a problem."
Protection of tenure
With no objective measure of success, firing a tenured teacher for incompetence is nearly impossible in many districts.
Earlier this year, a Times investigation found that jettisoning a tenured teacher solely because he or she can't teach is rare. In 80% of terminations upheld by the state in the last 15 years, classroom performance was not even a factor.
"Allowing ineffective teachers to remain in the classroom is literally dragging down the nation," said education researcher Eric Hanushek of Stanford University's Hoover Institution.
In a forthcoming paper, Hanushek estimates that replacing the nation's worst 6% to 10% of instructors with merely average teachers would propel the United States from its below-average level into the ranks of the world's top five educational systems.
So far, nobody is proposing even that. Most proponents hope to use value-added scores as one of several measures to identify teachers who need help.
Still, the potential of the value-added method has caught the eye of policy wonks, educational reformers and billionaire philanthropists such as Eli Broad, Bill Gates and brothers Lowell and Michael Milken.
But value-added has met with staunch opposition from teachers unions, which often argue that standardized tests are not a good measure of students' performance, let alone teachers'.
They also cite some experts' concerns that the statistical methods of the value-added approach may rest on ill-founded assumptions.
For example, some research shows it's easier to achieve gains with certain types of students. The result could be the unfair rewarding or punishing of teachers based on the students they are assigned.
"The danger is that too much weight will be put on this," said Helen Ladd, a Duke University professor, who urges more study. "I think the policymakers are ahead of the statisticians here."
But cracks in union opposition are beginning to show. Randi Weingarten, president of the 1.4-million-member American Federation of Teachers, once fought efforts to use the value-added approach in New York, and in August cautioned the Obama administration against "over-reliance on an unproven idea."
But this month, her union announced innovation awards for several local unions that plan to adopt value-added measures as part of a new teacher evaluation system even though Weingarten said she still has reservations.
"What we've realized is, if we don't do it by ourselves, no one is going to listen to us," she said.
An early skeptic
Grier was skeptical when he first encountered value-added results in the early 1990s as superintendent of schools in Williamson County, Tenn., just south of Nashville.
Though it's one of the wealthiest districts in the country, many students in its highest-achieving schools were not making much annual progress. Students in some of the poorer schools, meanwhile, were making remarkable gains, learning two years' worth of material in a single grade.
Then, at an academic conference, Grier was won over after seeing a presentation by William Sanders, then a University of Tennessee statistician.
Sanders described how he had developed a statistical tool that, in essence, took into account three years of a student's test scores to estimate the child's academic trajectory. Then he compared it with the student's actual performance in the fourth year. The difference between the expected growth and actual growth became known as "the teacher effect."
The approach overcame the Achilles' heel of traditional achievement tests, which critics say reflect socioeconomic status more than learning. They judge this year's students against last year's, ignoring potential differences between the two groups.
With the value-added method, students are compared to themselves from year to year, so the results are not skewed by income levels, parental involvement, race or gender.
"A teacher or principal has no responsibility for what's happened to the kids in the past," Sanders said. "What they do have responsibility for is the student's rate of progress. This is what was different about what we did."
Sanders' research found that there is a huge disparity in the effectiveness of teachers, and that good teachers could make an enormous difference in a student's learning, regardless of the child's background.
Subsequent research largely confirmed his findings, and in the process many orthodoxies of education policy were called into question. In particular, studies indicated that the traditional measures of a teacher's quality -- such as years of experience, credentials and education -- have little bearing on his or her effectiveness.
Grier was convinced. "I thought it could be a game-changer," he said.
To challenge students at his wealthy schools, Grier pushed principals to focus on growth rather than hitting fixed scores. He used the data to assign teachers to students who matched their strengths. He denied tenure to at least four teachers whose students did not show growth for several years, and reassigned some low-performing principals while encouraging others to retire.
At his next job in Guilford County, N.C., Grier hired Sanders to improve on the district's existing value-added program.
The goal of the program, dubbed Mission Possible, was to attract effective teachers to low-performing schools. To do this, Grier offered highly effective teachers up to $10,000 to relocate. They got up to $4,000 more if students' test scores rose.
Many educators protested the plan, but, with no collective bargaining in North Carolina, the board approved Mission Possible in 2006.
By 2007, teacher turnover at some of the area's lowest-performing schools had been cut by almost half, according to a recent independent study. But the program hasn't significantly changed student achievement -- at least, not yet.
"It's a mixed bag," said board member Alan Duncan. The real test, he said, will be a few years from now, when the board expects the changes to show up in test scores.
North Carolina has since adopted value-added statewide.
Grier said he knew that Mission Possible wasn't possible in San Diego. "The board made it clear they were not interested in a merit-pay program," he said.
Instead, Grier wanted to find out how the district's high-performing teachers were distributed. He suspected that many were clustered north of Interstate 8, in the wealthier part of the city.
He lured Sanders, now with a for-profit data analysis company, to visit San Diego with the promise of golf at Torrey Pines. The statistician's presentation persuaded the school board to approve a one-year, $80,000 contract with Sanders' company -- but limited its use to identifying students in need of extra help.
Union officials decried the move, saying the money would have been better spent hiring another teacher. They suspected Grier was trying to move the district toward merit pay. "We knew his history," said union President Camille Zombro.
She and other union members interrupted a school board meeting in June to present the board with a petition complaining about Grier's top-down management style and calling for his removal.
"He looked good in a suit, but he wasn't willing to collaborate," Zombro said.
Relations with the administrators union also soured when Grier tried to use value-added scores as a component of principal evaluation, a plan that was eventually rejected.
The board did not renew Sanders' contract this fall and never made public the results of his analysis.
Grier took over as Houston's superintendent in September. Reached by telephone recently, he seemed invigorated.
"They do things differently out here," he said. "It's a breath of fresh air."