How to Make Great Teachers

(3 of 5)
How Do You Measure Merit?
To the business-minded people who are increasingly running the nation's schools, there's an obvious solution to the problems of teacher quality and teacher turnover: offer better pay for better performance. The challenge is deciding who deserves the extra cash. Merit-pay movements in the 1920s, '50s and '80s stumbled over just that question, as the perception grew that bonuses were awarded to principals' pets. Charges of favoritism, along with unreliable funding and union opposition, sank such experiments.

But in an era when states are testing all students annually, there's a new, less subjective window onto how well a teacher does her job. As early as 1982, University of Tennessee statistician Sanders seized on the idea of using student test data to assess teacher performance. Working with elementary-school test results in Tennessee, he devised a way to calculate an individual teacher's contribution, or "value added," to student progress. Essentially, his method is this: he takes three or more years of student test results, projects a trajectory for each student based on past performance and then looks at whether, at the end of the year, the students in a given teacher's class tended to stay on course, soar above expectations or fall short. Sanders uses statistical methods to adjust for flaws and gaps in the data. "Under the best circumstances," he claims, "we can reliably identify the top 10% to 30% of teachers."

Sanders devised his method as a management tool for administrators, not necessarily as a basis for performance pay. But increasingly, that's what it is used for. Today he heads a group at the North Carolina–based software firm SAS, which performs value-added analysis for North Carolina, Ohio, Pennsylvania, Tennessee and districts in about 15 other states. Most use it to measure schoolwide performance, but some are beginning to use value-added calculations to determine bonuses for individual teachers.

Sanders' method is costly and complicated, however. Under steady pressure from NCLB to raise test scores, some districts have looked for quicker, easier ways to identify and reward teachers who boost achievement. In some cases, they have made the call largely on the basis of a single year's test results—a method experts dismiss as unreliable. In Florida, for instance, one of Governor Jeb Bush's final initiatives before he left office in January 2007 was to push through a merit-pay program that offered a 5% bonus to teachers in the top 25% in each participating district, with selection based at least 50% on how much their students' test scores jumped from one year to the next. Houston had a similar initiative, though without the 25% cap.

Both schemes met with fierce resistance. Teachers rebelled against the notion that a year's worth of instruction could be judged by how students did on a single test on a single day. They objected to the lack of clarity about how teachers of subjects not tested by the state would be assessed. And they railed against a system that pitted one colleague against another in a competition for bonuses. To make matters worse, there were gruesome glitches. In Houston, a newspaper website identified which teachers got bonuses. Later, 99 employees were asked to return about $74,000 in bonus checks issued by mistake. In Florida, one county ran short of bonus funds while another had an embarrassing discrepancy between the number of awards given in predominantly white schools and the number that went to schools with mainly black students. Both Florida and Houston have improved their programs, but local teachers remain wary. "The new plan doesn't have clear goals," charges Gayle Fallon, who heads the Houston Federation of Teachers. She fully expects "all hell to break loose again."

1. McCain's 7 Steps to Beating Obama
2. Obama: How He Learned to Win
3. Is It Time to Invade Burma?
4. The Five Mistakes Clinton Made
5. Toll Mounts in China Earthquake
6. Klein on Obama
7. Google Wants to Facebook Friend You
8. Study: 'Daily Show' A Lot Like O'Reilly
9. The Thriving Cult of Greed and Power
10. China: 18,000 'Buried' in One City