The debate -- and that’s putting it nicely -- over the use of standardized test scores in teacher evaluations has always confused me, because the answer seemed so simple. One of the things we ask of teachers -- but just one thing -- is to raise those scores. So they have some place in the evaluation. But how much? Easy. Get some good evidence and base the decisions on that, not on guessing. The quality of education is at stake, as well as people’s livelihoods.
Much to my surprise, at a meeting with the editorial board this week, Michelle Rhee agreed, more or less. As one of the more outspoken voices in the school-reform movement, Rhee is at least as polarizing as the topic of teacher evaluations, and her lobbying organization, Students First, takes the position that the standardized test scores of each teacher’s students should count for no less than 50% of that teacher’s rating on performance evaluations.
But asked where the evidence was to back up that or any other percentage figure, Rhee agreed quite openly that it’s lacking. The one robust study, funded by the Bill and Melinda Gates Foundation and released in January, suggested that test scores should count for a third to half of the evaluation. It found that scores were most closely associated with teacher quality as measured by various factors -- but they also were unreliable. In other words, the scores might go up or down in any given year, though they were quite useful over the long haul.
That’s just one study, though. As Rhee put it, it’s a good start. Not all the teachers went along with the parameters of the study, so the data on their results were less reliable. And the teachers weren’t facing high-stakes decisions about their jobs based on the test scores, which might also have an effect on the findings.
The problem is that Rhee slips in and out of her convictions on testing’s role. She concedes that other group think tests should count for 35% of the evaluation, and that the number might be anywhere between the two. She reasonably suggests that school districts try some different numbers as a sort of pilot program to see what seems to work best.
But what about trying less than 35%? At that point, Rhee refers back to the single study, saying it found problems if tests counted for less than a third. But that was minutes after she had agreed that a single study doesn’t give policymakers the evidence they need to determine what laws should be passed governing teacher evaluations.
And as head of the Washington, D.C., schools, Rhee counted the scores at 50% -- and that was before the Gates-funded study even came out. Well, you have to start somewhere, she said.
I suppose that’s true. But starting somewhere is a long way from making etched-in-stone decisions about teachers’ fates and how to bring about the best education possible in public-school classrooms. And that doesn’t just refer to Rhee. In response to President Obama’s demands of states that wanted Race to the Top funds, various states have imposed their own laws requiring test scores to count for up to 50% of evaluations. Though the president requires that the tests count for a “significant” portion, he never set a cap on what a reasonable portion would be. And what little evidence we have so far on scores suggests that they should be considered only as a trend over several years, not based on results from a year or two.
Maybe 50% is the right number. Maybe more than 50% is right -- or 15% would work better. I don’t know. The thing that’s so bothersome is that so many politicians, lobbyists and supposed experts are making hard-and-fast decisions when they don’t know the answer either.