Assessment Opinion

Should Teacher Evaluation Depend on Student Test Scores?

By Diane Ravitch — November 03, 2009 3 min read
  • Save to favorites
  • Print

Dear Deborah,

What a lovely tribute to Ted Sizer! I did not know Ted nearly as well as you did, but I admired him very much. He was very much the gentleman, and truly a gentle man. I had many disagreements with Jerry Bracey over the years; he was not gentle at all. Nonetheless, it is sad that these two men will no longer be among us, as they were both completely independent, a quality that is in short supply these days.

Which brings us back to the Obama agenda for education. Most educators are dubious about this agenda, but unwilling to speak up. The profession encourages timidity, I am sorry to say, because no one is supposed to speak out unless their supervisor approves, and the superintendents these days are looking at that big pile of cash in D.C. and hankering for a piece of it. So no one speaks up.

But that’s why we are here, so let’s have at it.

As you know, one of the big-ticket items on the Obama agenda is a proposal to evaluate teachers by looking at changes in their students’ test scores. As I explain in my forthcoming book, this idea comes out of studies by various economists who say that credentials and experience count for nothing, and that if we value improvements in student performance, we should judge teachers by their students’ scores. If the scores go up, the teacher is “effective,” and if they don’t go up, the teacher is a loser.

This approach has become wildly popular among the chattering classes. They think it is akin to a business that makes a profit (a winner) and one that loses money (a loser). They do not know of the studies by economists demonstrating that this particular measure of effectiveness is highly unstable. A teacher may have a class that gets higher scores one year, but not the next; or lower scores one year, but not the next. And then there is the fundamental problem, as all psychometricians warn us, that tests should be used for the purpose for which they were intended, and not for other purposes. In other words, a test of fifth grade reading tests whether students in the fifth grade are able to read material appropriate for children their age. It cannot then be used to determine whether their teacher was good or bad.

Writers who know nothing about education love the idea, however. For example, The New York Times published an editorial on Oct. 29 about the new teachers’ contract in New Haven, Conn., which will allow test scores to count when evaluating teachers. The Times was happy about that, but disappointed that the contract did not spell out a precise formula “in which the student achievement component carries the preponderance of the weight.” Instead, the details will be determined, to the Times’ chagrin, by a committee that includes teachers and administrators.

By coincidence, the Century Foundation published an issue brief by Gordon MacInnes on the same day titled “Eight Reasons Not to Tie Teacher Pay to Standardized Test Results.” Among the reasons are these: “Even reliable standardized tests are valid only when they are used for their intended purposes"; students are not randomly assigned to schools or to classes; state data systems are in their infancy, and it is far too soon to produce reliable and accurate longitudinal data; the assumption behind such plans is that teachers are holding back on their efforts because they are not paid enough (when it is far likelier that teachers, schools, and legislators “simply don’t know how to improve educational prospects for poor children”); such an approach will inhibit collaboration among teachers; and most teachers don’t teach a subject or grade that is subject to regular testing.

I have been trying to figure out how a school would function if the advocates of tying test scores to teacher evaluation prevail. At least three years of data would be needed, though five years would be better. At the end of the three-to-five years, the teachers who did not get gains would be fired and replaced by teachers who have no track record at all. Every year, a new group of teachers who had not produced gains would be fired, and another untested group of teachers would take their place. Most teachers, as MacInnes points out, would be exempt because they don’t teach reading or math. But for the unfortunate minority who do teach the tested subjects, there would be an annual game of musical chairs. There would be constant churn, with untried teachers thrown into the trenches. Some might make it (though it will take three years or more to be sure), but many will be ousted.

Does any other profession work this way?

Correct me if I am wrong, Deborah, but I don’t think this describes what any of the high-performing nations in the world do.


Related Tags:

The opinions expressed in Bridging Differences are strictly those of the author(s) and do not reflect the opinions or endorsement of Editorial Projects in Education, or any of its publications.