YOU ARE HERE: LAT HomeCollections

Student test scores can identify effective teachers, study finds

Teacher ratings are found to be more reliable when so-called value-added measurements are used along with other performance indicators

January 08, 2013|By Teresa Watanabe, Los Angeles Times

Student standardized test scores can accurately identify effective teachers, especially when combined with classroom observations and pupil surveys, according to a major national study released Tuesday.

The three-year survey of 3,000 teachers in seven school districts by the Bill & Melinda Gates Foundation found that the controversial method of measuring student academic growth, known as value-added, was a valid indicator of whether teachers helped boost student achievement. But the teacher rating was more reliable when test scores were used along with other performance measures.

Grappling with the key question of how much weight to place on test scores, the study recommended half to one-third of a teacher's evaluation.

The study also found that classroom observations should be performed by more than one person. And feedback from student surveys could provide a more accurate picture of a teacher's performance than occasional observations by educators.

As many school districts across the country overhaul their teacher evaluation systems, the Gates study may accelerate the use of standardized test scores as one measure to determine instructors' performance. The Obama administration has promoted that approach in its federal Race to the Top grants and other programs.

Tom Kane, a Harvard University professor of education and economics who led the $45-million Measures of Effective Teaching Project for the Gates group, said the results were the "strongest evidence so far" that value-added combined with other measures could accurately identify effective teachers. He emphasized that the project's purpose was to promote good teaching — not to nail instructors for poor performance.

"This is not about accountability," he said. "It's about providing feedback every professional needs to strive toward excellence. "

The report affirms the thrust of the Los Angeles Unified School District's new teacher evaluation system, said Supt. John Deasy. The system includes classroom observations, a student achievement measure and new pupil and parent surveys.

The district and United Teachers Los Angeles have forged a tentative agreement on the new system — which union members will vote on next week — that does not yet clarify how testing data will be used or how much it will count in the overall rating.

Deasy said the Gates report has "strengthened" his inclination toward counting test scores for about 30% of the evaluation, with observations making up the greatest share.

UTLA President Warren Fletcher, who has opposed value-added measures, noted that the study showed that using test scores for most of an evaluation made the results less reliable.

But the report also concluded that it would be "counterproductive" to count test scores for less than a quarter of a teacher's evaluation.

The value-added method, which calculates student academic growth after controlling for income, English ability and other factors outside an instructor's control, has been criticized by many teacher unions and some researchers. They argue that the method cannot accurately isolate a teacher's effect on student learning from other factors, such as disruptive classrooms.

But the Gates study found that teachers who were highly rated with their regular classes performed equally well when students were randomly assigned to them the following year.

"The report affirms that value-added is a useful tool — not alone, but in conjunction with other means of assessing teachers," said Dan Goldhaber, director of the Center for Education Data and Research at the University of Washington.

The results of the study are in line with the advocacy of the foundation, which has for several years funded groups that promote using value-added metrics as part of a teacher's evaluation.

But the study does not end the debate over value-added, said Jesse Rothstein, a UC Berkeley associate professor of public policy and economics. He said the project did not study whether evaluations used for high-stakes decisions will cause teachers to cheat, change what they teach or take other actions that could skew the results.

Morever, he said, a significant number of schools did not comply with the random assignment of students, reducing the number of teachers in the sample.

In a conference call, participants emphasized that their purpose was to provide research so school leaders could decide the best approach for their districts.

"We're not trying to settle a political debate," said Vicki Phillips of the Gates foundation. "We're trying to follow the evidence."

Times staff writer Howard Blume contributed to this report.

Los Angeles Times Articles