Is Supervising the Heck Out of Teachers the Answer?

By Kim Marshall

Two notions about teacher evaluation have the ring of truth: It’s important for principals to get into classrooms and observe, and teachers should be evaluated on how much their students learn. But both ideas can be implemented in ways that don’t improve teaching and fail to boost student achievement. Here’s how.

In most schools, teacher evaluation is something of a joke. “Oh, Ms. Jones, I see you’re being observed today,” says a colleague, noticing the power suit and dazzling earrings. A majority of evaluation visits are preannounced, and who can blame a nervous teacher for preparing a special lesson? Even when principals aren’t seeing a dog-and-pony show, their arrival in a classroom improves student behavior and skews what’s observed. Then there’s the problem of how little instruction busy principals see (one formal observation a year is the norm). High student achievement depends on first-rate instructional practices happening week after week, month after month. But are they? Principals make an educated guess about what’s happening during the 99.5 percent of the year when they’re not there, say a prayer—and rely on teachers’ professionalism.

As for evaluating teachers on student achievement, if the metric is standardized-test scores, there are practical and ethical difficulties: (a) The results of most tests are not available until summer—too late for May deadlines for teacher evaluations. (b) There are no standardized-test results for more than half of teachers, including those in art, music, physical education, and the primary grades. (c) Most tests are not designed to measure individual teachers, so it’s unfair to use them for evaluation. (d) Even the “value added” approach—measuring the gains students make from September to May—isn’t viable, since experts say three years of data are needed to make a fair judgment. (e) Many tests measure only lower-order skills and factual knowledge, so making them high-stakes will undermine high expectations. (f) Using test scores for evaluation could lead to more cheating by stressed-out teachers, who are, after all, the ones administering the tests. And (g) raising the stakes undermines the kind of collegiality that is essential to...