Standardized tests are a hot-button issue among teachers, raising concerns about how much time they must devote to prepping students as the stakes for the tests only get higher. So for many teachers, the idea of using students' scores to judge how well they teach – and to decide how much money they earn – is a nonstarter.
But a number of school districts use test scores, along with other measures, to do just that.
Often these systems reward all teachers in a school when students make certain gains. But increasingly, education reformers are pushing for a focus on how well an individual teacher does in improving the scores of his or her students over time.
When a district has the proper data, "value-added" formulas can do that while accounting for variables such as a child's economic background, advocates of these plans argue.
"Once we see the huge range in teachers' effectiveness [as research on test gains has shown], it doesn't make sense to leave that information off the table when we're trying to evaluate teacher performance," says Ross Wiener, a senior adviser at The Education Trust in Washington, which focuses on closing gaps in student achievement.
Assessments and methods for linking scores to teachers need improvement, Mr. Wiener acknowledges. But, he says, it's only fair to students "to get strategic about rewarding the best teachers and helping the weakest to improve," and removing those who don't improve after sustained assistance.
Opponents counter that student tests were never designed for evaluating teachers. Test-score analysis should be offered to teachers to help them pinpoint the needs of their students, says Randi Weingarten, president of the American Federation of Teachers. "But [teachers] aren't going to use it if it's viewed as a gotcha tool," she says.
As head of a teachers union in New York City, Ms. Weingarten negotiated a system that offered schoolwide bonuses in some cases based on test gains. But when it comes to decisions about individual pay, she says, no system fairly isolates the effects of teaching from other factors that influence student performance on tests.
Even various value-added models are not sufficient, she says. In Houston, for instance, "people have found that the pay results are not really aligned" with other indicators of who the best teachers are.
Still, momentum is growing for building more data systems and sophisticated formulas that advocates say can be part of the picture in fairly judging teachers. Sixteen states have the structures in place to create such measures, according to Data Quality Campaign, a group in Austin, Texas, that promotes the use of data for improving schools. And the federal stimulus package includes $250 million to help more states track such information.