Will value-added measurement survive the courts?

An ongoing argument raging across the country over whether student test score gains are a fair way to gauge a teacher’s skill has hit the courts.

Photo by Jess Loughborough/Flickr

Photo by Jess Loughborough/Flickr

In what may be among the first of many lawsuits over the new evaluations—which have been adopted by multiple states—the Florida teachers union is challenging the state’s use of test scores in decisions about which teachers are fired and which receive pay raises. The Florida Education Association argues the system violates the U.S. Constitution’s equal protection and due process clauses.

The debate over the new systems has often centered on the frequent errors in what’s known as value-added measurement, which can lead to effective teachers being misidentified as ineffective, and whether the potential problems for teachers outweigh the potential benefits for students.

A new paper published this week explores both sides.

Ratings for teachers based on test scores get it wrong a lot of the time. Dan Goldhaber of the University of Washington-Bothell, and Susanna Loeb, of Stanford University, review previous research that finds about a quarter of teachers are likely to be misidentified as ineffective when they’re in fact effective using the test score measures.

“The error rates,” they write, “appear to be quite high.”

And yet, the researchers argue that using test scores to make high-stakes decisions about teachers’ jobs is actually a more accurate method than previous systems, which often depended on cursory classroom observations, pass rates on licensure tests, and degrees earned.

“Flawed as they are, value-added measures appear to be better predictors of student achievement than the teacher characteristics that we currently use,” the researchers write.

What to do? Goldhaber and Loeb, as researchers are wont to do, suggest more research. But they also lay out some of the trade-offs that policymakers (and now the courts) should consider.

Is firing some teachers who may not deserve to be fired worth it if it ensures more bad teachers leave the classroom and more students have the opportunity to sit in the classroom of a good teacher?

Will increasing accountability, especially if the systems are viewed as unfair, have the unintended effect of reducing collaboration among teachers and scaring away good candidates from the profession?

Will the new evaluations prompt schools of education and districts to better train teachers, or will the newly-identified struggling teachers still be left to flounder?

Finally, the researchers ask, “How will the system handle legal challenges?”

At issue in Florida are not the error rates, but the fact that teachers are receiving ratings based on test scores of students or subjects they don’t teach. “This lawsuit highlights the absurdity of the evaluation system,” said Andy Ford, president of the Florida Education Association, in a statement.

Goldhaber and Loeb argue that for the most part, courts will probably defer to states and districts on the new evaluations, recognizing “that value-added measures are intended to improve those systems.” But they also say that the outcome of lawsuits will depend on whether “the state or district can demonstrate that the evaluation system in question was thoughtfully designed and consistent with sound educational principles.”

“Ultimately,” the researchers conclude, “employment decisions need only be based on evaluation systems that are sufficiently valid, not perfect.”


POSTED BY ON April 18, 2013